AVALI Business Opportunities from Space Technology. Technology Programme Report 8/2006 Final Report

AVALI Business Opportunities from Space Technology Technology Programme Report 8/2006 Final Report

AVALI Business Opportunities from Space Technology 2002 2005 Final Report Technology Programme Report 8/2006 Helsinki 2006

Tekes Your contact for Finnish Technology Tekes, the Finnish Funding Agency for Technology and Innovation, is the main funding organisation for applied and industrial R&D in Finland. Funding is granted from the state budget. Tekes primary objective is to promote the competitiveness of Finnish industry and the service sector by technological means. Activities aim to diversify production structures, increase production and exports and create a foundation for employment and social well-being. In 2006, Tekes will finance applied and industrial R&D in Finland to the extent of 460 million euros. The Tekes network in Finland and overseas offers excellent channels for cooperation with Finnish companies, universities and research institutes. Technology programmes part of the innovation chain Tekes technology programmes are an essential part of the Finnish innovation system. These programmes have proved to be an effective form of cooperation and networking for companies, universities and research institutes for developing innovative products, processes and services. Technology programmes boost development in specific sectors of technology or industry, and the results of the research work are passed on to business systematically. The programmes also serve as excellent frameworks for international R&D cooperation. Copyright Tekes 2006. All rights reserved. This publication includes materials protected under copyright law, the copyright for which is held by Tekes or a third party. The materials appearing in publications may not be used for commercial purposes. The contents of publications are the opinion of the writers and do not represent the official position of Tekes. Tekes bears no responsibility for any possible damages arising from their use. The original source must be mentioned when quoting from the materials. ISSN 1239-1336 ISBN 952-457-230-3 Cover: Oddball Graphics Oy Page layout: DTPage Oy Printers: Libris Oy, 2006

Foreword The emphasis of the Finnish national space strategy is on diverse exploitation in Finnish society of the opportunities provided by space technology. The goal is for research and development in the space sector to enhance service production and industrial competitiveness. The technology programme AVALI Business Opportunities from Space Technology 2002 2005 has met this need. Preparations for the programme revealed a need in Finland for development of the technological basis for business which utilises space technology. There had been little business that exploited space technology, and players in the sector were interested in a development framework for the satellite sector, remote sensing, and other related applications. The technological basis for business already existed. The potential for business opportunities was supported in particular by the high level of Finnish competence in satellite and airborne remote sensing and by the rapidly growing telecommunications industry and the research on which it is based. Moreover, new satellite systems such as the European Galileo satellite navigation system were also in the offing. During preparations for the AVALI programme, it became apparent that demand for commercial satellites would increase in the future. It was expected that this would also increase the opportunities for Finns in commercial operations outside the European Space Agency (ESA). This did not, however, prove to be the case because the recession that hit the telecommunications industry at the outset of the programme also had an impact on the market for commercial satellites. Major satellite suppliers were forced to cut costs and consolidate. This in turn affected orders in both ESA and commercial subcontracting, although the situation was also seen as a longer-term opportunity to gain access to supplier networks. The focus of the programme expanded. The overriding goal was to create a foundation for business that would benefit space technology and infrastructure in one way or another. Hence, hardware and software for the satellite and terrestrial segment and the development of commercial remote sensing, for example satellite remote sensing, airborne remote sensing and radar technology, were part of the AVALI programme. Moreover, exploitation of the technology and competence developed in the space projects was chosen as one of the priorities of the programme in external space-sector applications. Research and development projects formed the backbone of the AVALI programme, the extent of which grew from the projected 15M to more than 22M. Altogether 35 companies and 22 research or other public organisations

participated in the programme. In the end, the projects were mainly related to development of remote sensing applications and hardware and to building of the foundations necessary for business in these areas. The number and quality of the spin-off projects were also significant. The programme offered a framework for development of new kinds of products that exploit space technology. The number of development projects aimed directly at the satellite sector turned out to be quite small. In addition to the projects, the programme also collected data on commercial markets and business prospects. It created links between Finnish suppliers and international customers through both direct contacts and cooperative endeavours. The cooperation agreement between Tekes and the Canadian Space Agency, which provided a framework for initiation of joint projects aimed at the development of remote sensing applications, was of special significance. Apart from the results of the projects, it is obvious that the AVALI programme influenced the thinking of Finnish space sector players and users and made it more business-oriented. Many individuals contributed to the success of the AVALI programme. Tekes would like to thank the Management Board of the programme and its chairman for their active role in the programme and in steering the actions taken. The extensive expertise of the Group and the discussions held in its meetings were a valuable addition to the content of the programme. Our special gratitude goes to AVALI Programme Manager Pekka Havola of the Turku Science Park Oy for his experience and determination in implementing the ideas of the Management Group. Helsinki, March 2006 Tekes, the Finnish Funding Agency for Technology and Innovation

Contents Foreword 1 The AVALI programme....1 1.1 Background...1 1.2 The key themes...2 1.3 Objectives of the AVALI programme...2 1.4 The Management Board and the operational structure of the programme....3 1.5 Midterm review of the programme...4 2 Activities to create business opportunities for Finnish space industry....5 2.1 International activities....5 2.2 National activities...8 3 AVALI projects...11 3.1 Project criteria...11 3.2 Facts and figures...11 3.3 Company projects...13 Market Analysis Commercial Space... 13 TerraSAR-X Leaf Amplifier Assembly....13 SPAXMON Market Study...14 Development of Power Control and Distribution unit (PCDU)...15 Modular Power Supply Development (MPWR)... 16 Improvement of impulse radar (ImRa)... 17 Hyperspectral Imaging Technology...18 ClusterMaster...19 A Service for defining forest patterns...20 Automatic tie points for digital aerial images...20 Development of a Pseudolite System...21 Modular extensions for ESAComp software....22 FDIR Applications...23 Business from Millimeter Waves (MILLI)...23 Asynchronous Test Access Port...24 3.4 Research projects....25 Environment monitoring using Earth Observation (ENVIMON)... 25 Advanced Techniques for Mobile Positioning (MOT).... 28 Improving the EGNOS Service in Finland with Internet Radio Technology (IESIR)... 30 Development of the RADEF-irradiation station...31 Commercialization of the on-wafer test services of MilliLab...32

3.5 Consortium projects...33 Dual polarization Doppler weather radar...33 Oil spill detection information in the Baltic Sea region (OILI)... 39 Use of individual tree interpretation obtained from aerial images and laser scanner further development of methods and market analysis....43 Forest Parameter Estimation from High Resolution Remote Sensing Data...46 Quality of laser scanning, especially in urban environments (LAQU)... 53 Statistically calibrated satellite image interpretation for land use and forest characteristics...58 Interpretation methods for waveform laser scanning (Waveform)... 60 Helsinki Testbed...61 GaAs Imaging X-ray Detectors...64 3.6 Finnish-Canadian cooperation projects...68 Development and Demonstration of Remote Sensing Based Products For Surface Water Management (WRPD)...68 Northern Boreal Forest Information Products Based on Earth Observation Data...71 Applications and software for SAR interferometry and coherent target monitoring (AppliSARIN)... 73 Multi-Polarisation SAR for Operational Sea Ice Monitoring (POL-ICE)... 76 Attachments Companies....79 Research Organisations...79 Tekes Technology Programme Reports in English....80

1 The AVALI Programme 1.1 Background Finland entered the space sector in the mid-1980s through bilateral agreements with neighbouring countries. In 1987, Finland became an Associate Member of the European Space Agency. Full membership followed in 1995. Finnish scientists and industry have played an important role in many European satellite missions including ESA s SOHO and Cluster, ENVISAT, MSG, and METOP. More recent bilateral programmes have involved NASA s EOS-Aura and Sweden s Odin satellites. In addition to ESA space science and earth observation programmes, Finland has joined the European large telecommunications platform effort AlphaBus and the Galileo satellite navigation programme. Wide utilisation of earth observation data in operative environmental applications and environmental research is one of Finland s key competencies, and provides an excellent basis for the country s investments in space. Space sector activity is broad-based. There are a number of user organisations on many levels from the sciences to content production and the entertainment industry. In Finland, space technology offers an innovative R&D environment and provides opportunities for creating contacts with the R&D activity of major international companies, research institutes, and small innovative companies. Focus areas of Finnish space activity are remote sensing space science satellite telecommunication satellite navigation manufacturing of space equipment. Figure 1. Focus areas of Finnish space activity. 1

Many Finnish companies are technologically well equipped to work directly in the space segment market or as subcontractors to companies operating there. Although expertise in the application area is strong in Finland, commercial activity has so far been small compared to institutional activity. Nevertheless, the application market offers a socially and commercially significant opportunity. Tekes has had two technology programmes related to the space sector: Space 2000 Space Equipment Technology and Globe 2000 Remote Sensing Technology. Both programmes ended in 2000. These were followed by the Antares Space research programme from 2001 to 2004, which was conducted and finalised in collaboration with the Academy of Finland. The Finnish national space strategy (Ministry of Trade and Industry Advisory Committee Report 1/1999) defines the objectives and strategy for development of space activity. One key issue that it refers to is the expansion of business activity to the market outside ESA. The strategy also requires that the use of space technology be extended in other areas of industry. The next space strategy was approved at the beginning of 2002 and thereafter in 2005, both in accordance with the earlier strategies. Applications based upon spaceborne data and satellites The focus is upon remote sensing applications and products together with the application of satellite infrastructure such as telecom and navigation technologies. Further to this, preparatory activities designed to create business opportunities are also actively encouraged. In some cases an operative system may have to be developed prior to the commencing of a commercial application. Airborne remote sensing was also included in this theme. Satellite hardware and software Satellite subsystems, equipment, components, and software are key areas of interest, particularly potential spin-offs from the terrestrial industry to the space industry. The commercialisation of oneoff space developments and niche products will also provide interesting potential for future space-related business Ground Segment hardware and software Finally, equipment, software, and services for the realisation of satellite systems and for the direct processing of satellite data would also constitute a good basis for valuable co-operation. 1.2 The key themes AVALI was a business-oriented technology programme launched by Tekes in 2002 and finalized in 2005. It was conceived with the intention of creating new business opportunities, products, and services based on space infrastructure and know-how created in space projects. Such applications are designed to enable both new services and product concepts. The AVALI programme focused on the following four key themes: Utilisation of space technology for terrestrial products and applications The utilisation of spaceborne knowledge and technology in terrestrial products and services is of specific interest to AVALI. 1.3 Objectives of the AVALI programme The vision of space business activity was defined by the AVALI programme at 2002 as follows when the programme plan was launched. Finnish companies will have continuous and profitable business activity in satellite segment products and in the application segment. Companies will form part of the delivery chains of the main European and non-european prime contractor companies and equipment manufacturers. The service products supplied by Finns will be in demand. The main focus of business activity will be on the market outside ESA s industrial return regulations and nationally financed activity. Tech- 2

nological development will be supported by ESA activity and in national projects. Companies and research institutes will transfer the technology and expertise from space projects for terrestrial use in business. The overriding goal of the AVALI programme in accordance with the vision The goal was to create new and viable business on the commercial market with space technology, applications of it, and services. The main focus of the business activity was on the market outside ESA s industrial return regulations and nationally funded activity. The development of systems and equipment for commercial satellite markets represented a key area within the programme and the main goals of the programme can be expressed as follows: to increase the competitiveness of the terrestrial products and applications of the companies with space technology to create the foundation for commercial applications implemented with space infrastructure to develop products for the commercial satellite market. 1.4 The Management Board and the operational structure of the programme The AVALI Management Board was the main body providing guidelines and strategic issues for the programme. The Management Board also had a very important role in selecting and proposing the tools with which the AVALI programme would create the conditions for commercialisation of Finnish space technologies, services, and products. Tekes invited individual persons with long experience of space-related business, research activities, or ESA projects to become Management Board members. Their experience covered all the key themes of the AVALI programme. Management Board operations The main issues dealt with in Management Board meetings were the initiations, proposals, and approvals of the tools with which the AVALI programme would create the conditions for commercialisation of Finnish space technologies, services, and products during the programme. The proposed activities were preliminary studies or surveys of common interest, project specific market studies and reports, definitions of customer requirements and product specifications, and joint market surveys. AVALI Management Board Tekes Programme Manager AVALI Programme Manager PROJECTS Research projects Company projects Consortium Projects Finnish-Canadian cooperation projects OPERATIONAL ACTIVITIES Preliminary studies Market surveys Surveys of common interest Seminars and Workshops Figure 2. Operational structure of AVALI programme. 3

The Management Board members played an active role in proposing international speakers for seminars and in creating the content/programme for seminars and workshops concentrating on themes relevant to the programme. The Management Board was not involved in funding decisions and did not make any proposals or decisions concerning research or company projects. The Tekes Programme Manager and the AVALI Programme Manager played a significant role in daily programme management. The structure of the programme is presented in figure 2. The Management Board had six members Chief Scientist Heikki Sipilä of Oxford Instruments Analytical Oy, chairman of the AVALI Management Board General Manager Business Development, Juhani Hanka, Patria Aerostructures Research Professor Tuomas Häme, the Technical Research Centre of Finland (VTT) Director Fredrick von Schoultz, Stok - Development Centre (formerly worked at Space Systems Finland Oy) Director Miranda Saarentaus, Soil and Water Oy, Jaakko Pöyry Infra Chief Technology Adviser Heikki Hannula, Tekes Pekka Havola, Turku Science Park Oy, was programme manager and secretary to the Management Board. The Management Board held 17 meetings and attendance at meetings was very close to 100%. 1.5 Midterm review of the programme The midterm programme review was conducted by interviewing Management Board members. The work was conducted by LTT-Research Ltd., a company established by the Helsinki School of Economics. The evaluation was based on the material of the AVALI programme such as brochures, minutes of Management Board meetings, reports, and on a questionnaire directed to Management Board members. On completion of the midterm review, the conclusions and recommendations below were made. The midterm Review showed that implementation was in line with the goals of the programme. Focusing the programme With the ultimate effectiveness of the programme in mind, will the programme reach the goals set for it with the present broad focus, or should priorities be altered? Is the programme too splintered and the projects too small? Should the focus be on what works or on what doesn t? Taking value networks into account, business goals would be more readily attained if the players in the knowledge- and business-related value networks that are linked to innovations and their subsequent development paths were more involved in the programme. the criteria for access to the subcontractor chains in the space business should be defined. a report on the value / business chains for remote sensing should improve allocation of input/contributions and increase effectiveness. Building direct contacts More effort is needed to promote access to networks, build contacts between companies, and form partnerships. Effort should be put into the co-operation with Canada no such opportunities exist in other sectors / programmes. Completion of the North American market research project is important attention should be focused on establishing the proper focus when building partnerships. Overall, more resources in business-oriented technology programmes should be allocated to collecting a large amount of commercial information and to building networks and contacts. The evaluation also focused on what the role of Tekes should be and on what the added value of funding is when development issues are very close to the markets and to the business of the companies. 4

2 Activities to create business opportunities for Finnish space industry This chapter describes the activities that were proposed by the Management Board and implemented in the programme. These activities were proposed or initiated by the Management Board, Tekes, or the AVALI Programme Manager. The results will be used by all organisations participating in the AVALI -programme and also by other organisations outside of the programme. 2.1 International activities World Space Congress and Exhibition 2002 Tekes had decided that the first step in the AVALI programme would be to organise a joint stand for Finnish companies at WSC 2002. The world space community meets once every decade for a major congress and exhibition, which was held this time from 10 th to 19 th October in Houston, USA. The aim of Finnish participation would be to obtain visibility, make personal contacts with potential commercial customers, and discuss co-operation regarding sales of Finnish technology on the global market. The following took part in the exhibition: Aboa Space Research Oy, Space Systems Finland Oy, Oxford Instruments Analytical Oy (the former Metorex International Oy), Soil and Water Oy, Jaakko Pöyry Infra (the former Novosat Oy), Radef Research Laboratory, Patria Systems Oy (the former Patria New Technologies Oy ja Patria Finavicomp Oy), Componeering Oy, Millilab, Invers Oy, Elektrobit Microwave Oy (the former Ylinen Electronics Oy), Tekes. Turku Science Park Oy (the former Turun Teknologiakeskus Oy) was in charger of the practical arrangements for the exhibition together with Tekes. Some thirty Finns were at the exhibition, answering the questions of visitors. Figure 3. The Finnish stand at the 2002 exhibition. 5

Figure 4. Finnish representatives at the exhibition. Some of the participants immediately pointed out that the presence of their company had been useful. On the basis of the contacts made, the business opportunities exist and contacts since the exhibition have been close. There was obvious interest in the Finnish stand. Opportunities for Finnish Companies in the North American Space Market A market study North American Market Opportunities for Finnish Companies engaged in Space Business Activities was conducted during 2004 and 2005 as part of the AVALI programme. The study was ordered from International Space Services, Inc. after the invitation to tender. The aim of the market study was to provide background information for programme participants. The United States and Canadian space market generated approximately 42 billion U.S. dollars in revenue during 2003. Funding for North American space projects has been increasing for the past three years. US/NA is the leading space market in the world." The United States has also established the world s largest civil and military space programmes and funds a significant share of the world s space technology development. These projects will be a potentially attractive market for Finnish space firms. The study concentrated in the commercially most potential Finnish space segment technologies and on the Finnish technologies directly supporting them. This included the following areas in particular: Spacecraft structures and structural elements Satellite power systems Onboard satellite data processing systems Onboard software Detectors and detector systems including x-ray detectors RF-technologies and components. During the study period, the authors interviewed representatives in nine well-known North American companies concerning Finnish space products and the services offered by the 11 Finnish companies participating in the study. Overall, the study concluded that the Finnish companies have strong technical capabilities that can give them a competitive advantage in the in- 6

ternational marketplace. Further, the excellent reputations of several Finnish high technology non-space firms have generally established the country as a reliable, competitive supplier of high technology products and services. For space companies and service providers in Finland, this potential market in North America is a highly attractive one. For many Finnish firms seeking to enter the North American market for the first time the direct sales approach will be the most effective one. Nevertheless, there are other possible approaches including forming marketing alliances with North American firms and establishing U.S. or Canadian subsidiaries. Partnering is also a way to become known to North American customers. Successful marketing involves a wide variety of activities. It also involves developing close and ongoing relationships with potential customers and remembering the human touch. The study showed that commercial satellite manufacturers seeking new suppliers are also ready to include Finnish subcontractors in their supply chains. Finnish firms with technological capabilities should consider encouraging their key technical personnel to join and actively participate in the meetings of U.S. professional societies such as the American Institute of Aeronautics and Astronautics (AIAA) and the American Astronautical Society (AAS). Membership in the Institute of Electronics and Electrical Engineering (IEEE), and the International Society for Optical Engineering (SPIE) would be beneficial. Price is one of the most important competitive characteristics in today s market, though for certain scientific or military spacecraft, extreme reliability or special technical capabilities may be more important. The Traffic in Arms Regulations (ITAR) and the Missile Technology Control Regime (MTCR) are significant and must be considered before firms enter the North American space market. Finnish Canadian cooperation The basic idea behind the cooperation was to find solutions from space activity for the problems of northern regions. Thanks to their geographical locations, Finland and Canada have a lot in common. In May 2003, Tekes and the CSA (the Canadian Space Agency) signed a memorandum of understanding document regarding research and development in the space sector. Call for projects proposals concerning cooperation in remote sensing applications between Finland and Canada began during the visit of Marc Garneau, President of the Canadian Space Agency, to Finland. Thanks to this co-operation, Finns gained better access to measurements taken for example by Canadian RadarSat satellites. The Canadian radar satellites serve Finnish needs effectively because they also convey information during dark and cloudy periods in winter. With satellite images, for example, oil transports on the Baltic Sea can be monitored continuously and less expensively than by having ships photographed from an aircraft, as has been the case to date. For their part, the Canadians were motivated by Finnish expertise in research on the environment of northern regions. These common interests were successful in motivating applicants, and 20 participants expressed their interest during the first round. The projects concentrate on northern issues: snow, ice, forests, and the sea. Results can be expected in the form of improved weather forecasts in the future, inventory methods for forest owners, charting of floods, and ice services for shipping. The goal of co-operation is more effective use of remote sensing in northern conditions and attainment of world leadership in the development of remote sensing technologies. It is hoped that joint projects involving research institutes and industry will generate new business in accordance with the principal aims of the AVALI programme. In these projects both the Finns and the Canadians will fund their own participation in accordance with the practices prevalent in their countries. Tekes will contribute to the funding of Finnish efforts together with companies and research institutes. Short descriptions of the co-operation projects are presented in section 3.6 of this report. Joint Finnish-Canadian projects will continue, even after the AVALI programme. 7

EU Satellite Centre The function of the European Union Satellite Centre (EUSC) is to generate data for the Council of the European Union and for the EU Member States to support decision-making. The EUSC acquires satellite and other data, processes and analyses them, and prepares reports for EU decision-makers The EUSC is located in Torrejón, Spain, just outside Madrid. The EUSC would offer numerous opportunities and constitute an excellent reference for Finnish companies. With the assistance of Finpro, Tekes arranged a visit to the EUSC in May 2005 within the framework of the AVALI programme. The aim was to present those companies and organisations with something to offer regarding the remote sensing applications.. Representatives from five companies, one research institute, Tekes, and Finpro took part in the visit. 2.2 National activities The Finnish remote sensing survey More than half of the projects in the AVALI programme are related either to remote sensing or to its applications and equipment. After the second seminar of the AVALI programme, discussions in the Management Board indicated that there was not enough information about business opportunities in the remote sensing sector. The Management Board recommended that a survey of the sector be made. The survey concerned actors, products and services, service providers, customers, operational chains, the extent of commercial operations, and the outlook for the future. The goal would be a summary of the sector s actors, volumes of operations, present operational modes and supplier chains, and of customers as well. A key objective was to seek out areas in which there could be potential for commercial operations. Future prospects were to be estimated over a 5 to 10 year period. Similar needs occasioned by the geographical location gave cause to compare remote sensing carried out by Finland and Canada. The results of the report were intended for use by all Finnish organisations of remote sensing. For purposes of the report, remote sensing was defined as that made from satellites, aircraft, helicopters, or other airborne devices. After competitive bidding, Soil and Water Oy prepared the report. The study was made in two parts. First, an extensive written questionnaire was sent to players in the sector. Then personal interviews were made either by telephone or in personal meetings. The questionnaire was sent to 191 players and 67 responses were received (35%). On the basis of the information obtained, net sales from Finnish remote sensing amounted to 27M /year, including public funding. Private funding accounted for 61%. The questionnaire provided valuable information on the structure of the markets and a picture of the roles of players in the operational chain. Figure 5 presents the roles of players in data supply chain. The report concluded that there is a need for demonstrating the benefits of remote sensing, for making definitions, for projects to exploit existing data, and for various kinds of projects related to the distribution of data. It is also noteworthy that the distribution of remote sensing images via data networks is becoming widespread. Institutional organisations have a very high level of competence in the sector, which means that they both show the way to others and are also the first users of many applications. Research institutes should play an active role in seeing that their expertise is transferred and used in companies and exports. Joint projects involving research institutes and companies are recommended for the transfer of knowledge to the latter. Exchanges of personnel between companies and research institutes are also recommended for the transfer of knowledge to commercial operations.. It is hoped that institutional production and services could be outsourced commercially. It is also hoped that the prices of data will decrease and availability increase, thereby making use more widespread and increasing efficiency. 8

Research/development (products, software) Producer Data processor Service End user Data (primary production) Transfer and delivery Hardware Basic processing Potential own use Data processing (classification ec.) End product Demonstrations Software, applications Sales Company/Institute Personal customer Figure 5. The data supply chain in remote sensing. Source: Kaukokartoitusselvitys, Teknologiakatsaus 186/2005, Tekes. Research / Development 18 % End User 16 % Primary Producer 12 % Data processors 25 % Service Producers 29 % Figure 6. Percentages of questionnaire respondents by role. Source: Kaukokartoitusselvitys, Teknologiakatsaus 186/2005, Tekes. According to the report, the market for remote sensing in Finland is rather small and dependent on research. Consequently, business cannot be generated on the basis of the Finnish market alone. Excellence exists, but usually in government agencies. Data processors and service producers account for some 45% in institutional organisations. In private companies the figure is approximately 65%. The number of remote sensing personnel in companies was two in more than half of the responding companies. Several groups of ten were found, but there were only a few groups of more than ten. Particularly in primary production and service production, operations are commercial. Two-thirds of the service producers are private and they offer image processing, consulting, aerial photography, and research services. The products and services are diverse, even though aerial photography has long dominated. Commercial software products are limited and they are usually linked to instruments. Ninety-five per cent of the end users are institutional organisations. 9

AVALI Seminars and Workshops The main communication channels of the programme were annual seminars and other meetings. All seminar and workshop presentations were published on the AVALI website. E-mail lists were used also for promoting AVALI seminars and events. Additional information on progress in the AVALI programme was presented at the Industrial meetings of Tekes. Promotion of all major events was improved by advertisements in the technological newspaper Tekniikka & Talous. The first AVALI seminar was held on 22 nd January 2003 The main topics were the global space market situation and future forecast, the CSA s commercial activities, and the AVALI programme. The first annual seminar was held on 26 th May 2004 The main theme was the presentations of AVALI projects and especially success stories concerning commercialisation of project results. The AVALI programme arranged three workshops. Remote Sensing workshop on 3 rd September 2003 The objective of the workshop was to provide potential users with information on remote sensing and its commercial potential. The invitation was sent to some 150 organisations in the following sectors: insurance, banking, real estate, construction, transport, shipping, and security and multisector firms and telecommunications companies. Detection and measurement in the process industry on 6 th September 2004 and Detection and measurement in the environmental industry on 13 th September 2004 The AVALI programme took the initiative in arranging seminars for space equipment to exploit the measurement technology developed in the industry. These events were arranged together with other Tekes programmes. The second annual seminar was held on 1 st September 2005 The ongoing theme was commercialisation. Most of the presentations dealt with remote sensing applications. The final seminar was held on 4 th April 2006 10

3 AVALI projects 3.1 Project criteria The AVALI Management board had a lengthy discussion about in the programme s project criteria at the beginning of the programme. The following additional criteria were proposed: 1. Markets An assessment of the markets, the business sought, and business development should be presented. 2. Market orientation and global markets Does enough market volume exist? 3. Projects aimed at commercial space activity Applications must state the market segment sought and give an account of the prevailing market situation. 4. Spin-off products and services Spin-off products can seek markets that the company does not know well. How should one proceed in the market area?. 5. What are the trends, changes and discontinuities? 6. Customer needs What will customers need in the future, and particularly when the project is commercialised? 9. Research projects and co-operation between companies How has expertise in Finland and abroad been taken into account? 3.2 Facts and figures The total programme volume was 22.3 M million of wich 8.9 M for research projects and 13.4 M for industrial projects. Total Tekes financing for programme was 11.4 M. It consisted of 6.3 M for research and of 5.1 M for industrial projects. The total number of projects was 61. Details concerning application areas and finance are presented in figures 7 and 8. AVALI projects 2002 2005 Area No. Costs/M Space product 5 1.1 Remote sensing device 5 7.7 Remote sensing application 37 8.8 Other application 5 3.2 Service product 2 0.4 Spin-off 7 1.1 Total 61 22.3 7. Competitors and competitive advantages Who are the competitors, what are they making, and how has this been determined? 8. Business chain Applications should be prioritised in terms of how the entire business chain has been taken into account in the project (joint projects). A partner who knows the sector is an advantage when applying. 11

Service product 2% Other application 15% Spin-off 5% Space product 5% Remote sensing device 34% Remote sensing application 39% Figure 7. AVALI projects 2002 2005 by volume. Other application 8% Service product 3% Spin-off 11% Space product 8% Remote sensing device 8% Remote sensing application 62% Figure 8. AVALI projects 2002 2005 by number of projects. 12

3.3 Company projects Market Analysis Commercial Space Project objectives The objective of the Market Analysis Commercial Space project was to analyze the commercial space market mainly in Europe, the USA and Canada to provide a basis for establishing business in carefully selected target markets. Activities The activities consisted of commercial market segmentation in which the potential areas where analyzed. The starting point was the existing product and service base of the company. A detailed market analysis of each segment was made. It was based on various reports generated by dedicated consulting companies and research organizations and on direct contacts with different actors in the market. Project duration: from 1 st May 2002 to 15 th January 2004. Results The project provided important information about the state of the commercial market in selected markets. The markets are dominated by a set of large companies that use subcontractors in dedicated product/service areas. Our analysis showed that there are opportunities for smaller companies to participate in selected niche areas. This will require long-term product/service development in order to provide competitive products and services. The financing of the development has to be long-term and the risks are high. Project participants Space Systems Finland Oy Contact information Space Systems Finland Oy Sami Laitinen, project manager Tel. +358 9 613 28 623 sami.laitinen@ssf.fi http://www.ssf.fi TerraSAR-X Leaf Amplifier Assembly Background The project was started to create co-operation with a large space hardware manufacturer. The expertise of Elektrobit Microwave Ltd. was determined to be appropriate for manufacture of the required microwave unit. Project objectives The objectives were to design, manufacture and test a transmitter power divider/receiver combiner network for the microwave Synthetic Aperture Radar (SAR) satellite. Another objective was to enter the commercial space hardware manufacturing business. Activities One engineering model and two flight models of the unit were manufactured. Project duration: from 1 st October 2002 to 30 th June 2005. Results The required hardware was successfully manufactured. Some specifications were modified during the project to fulfil customers needs. New manufacturing technologies were adopted. Commercial impact Contacts to a customer were created. A good opportunity to gain access to subsequent SAR satellite projects was achieved. Some additional component delivery contracts were signed during the project. Project participants The project partner was EADS Astrium GmbH. Contact information Elektrobit Microwave Ltd. Petri Jukkala, project manager Tel. +358 9 57121723 petri.jukkala@elektrobit.com www.elektrobit.com 13

SPAXMON Market Study Background The impact of space weather is expected to increase in the near future. Technological systems in both space and on the ground have increased in number and also become more sensitive to consequences of interaction between ionising radiation and physical materials. Based on previous studies and space equipment development, a new instrument concept called Standard Radiation and X-ray Monitor (SPAXMON) was developed. This instrument is intended for measurement of the ionising radiation environment in space conditions. detector and several surface detectors. The surface detectors are pointed in different directions, an arrangement that also provides directional information on the incidence angle of the radiation. Preliminary estimates for the sensitivity of the detectors and the associated electronics were made. Solid-state detector energy ranges of 0.5 30 MeV were estimated for protons, 100 kev 3 MeV for electrons (optional), and 2.6 15 kev for X-rays. For gas proportional counter detector energy ranges of > 3 MeV for protons, > 300 kev for electrons and 1.5 30 kev for X-rays are the baseline. Project objectives The essential feature of the proposed SPAXMON instrument concept is that simultaneous charge particle flux and X-ray intensity could be measured with the same instrument detector. This feature is considered valuable when the power consumption, mass and volume of the measurement instrument need to be minimized. The underlying aim is to develop a compact instrument that both generates useful information on the space environment and is commercially attractive, so that satellite builders can easily integrate it with their future satellite missions. Activities The SPAXMON concept was defined on the basis of earlier X-ray measurement instrumentation. The instrument is specified for the detection of protons and X-ray photons. Electron detection capability is optional. A market survey was performed using a questionnaire form that was sent to experts in the field of space weather. Personal interviews were also made to obtain information on the commercial potential of the instrument. Project duration: from 1 st January 2004 to 31 st March 2005. Results The instrument s main component, a compound semiconductor detector, is composed of a core Figure 1. SPAXMON equipped with a solid-state detector. Commercial impact According to the market survey, the commercial potential of SPAXMON is estimated to be roughly 2 units per year, mainly to European space missions. However, achieving this potential would require further demonstration of performance. In the field of radiation measurements there is also considerable competition in Europe. Integration of the SPAXMON detector with only a larger space environment analysis system could be an interesting option for commercialisation. Project participants Patria Systems Oy Oxford Instruments Analytical Oy Observatory, Helsinki University Finnish Meteorological Institute. 14

Contact information Patria Systems Oy www.patria.fi Commercial matters Kimmo Myllyoja, director, space Tel. +358 20 469 2325 kimmo.myllyoja@patria.fi Technical matters Jari Stenberg, project manager Tel. +358 20 469 2335 jari.stenberg@patria.fi Development of Power Control and Distribution unit (PCDU) Background To a large extent, Patria space business focuses on power equipment and systems. Past experience includes satellite power distribution units and low power DC/DC converters. The tendency of putting the power control unit (PCU) and the power distribution unit (PDU) into a single power control and distribution unit (PCDU) was the main reason for the PCDU development project. Project objectives The primary project objective in the AVALI program was to design and develop a competitive overall power system between the power sources, the solar arrays and battery, and the satellite power consumers, thereby meeting the design requirements and targets with specific power demands. The equipment was divided into three main sections: 1) the PCU section, which provides solar array power (2.5kW) through sequential switching shunt regulators (S3R) to an unregulated power bus, 2) the DCDC converter section, which provides a reliable1 kw regulated power bus for users, and 3) the PDU section, which provides high current latching current limiters (LCL). Activities The project started with an analysis of the system requirements and the trade-offs of potential implementations. The final system concept consists of the bread board modules (2 S3R modules, 4 DCDC modules and one LCLU module) as presented in the figure 1 below. The unit was also tested according to the set requirements in hot and cold temperatures at the Figure 1. 15

Patria thermal chamber. The relevant test equipment, battery simulator and load simulator were developed and manufactured. A solar array simulator was acquired as part of Patria own investments. Project duration: from 1 st December 2003 to 31 st May 2005 Results The system that was developed achieved the design requirements and targets. Commercial impact The project achieved its commercial objectives. The design can be referred to and reused in future space programmes. Project participants Patria Systems Oy Tampere University of Technology EADS Astrium Ltd. Contact information Patria Systems Oy, www.patria.fi Commercial matters Kimmo Myllyoja, director, space Tel. +358 20 469 2325 kimmo.myllyoja@patria.fi Technical matters Juhani Simola, project manager juhani.simola@patria.fi Tel. +358 20 469 2333 Modular Power Supply Development (MPWR) Background Space programme applications have widely varying needs regarding power supplies and their performance. Nevertheless, all power supply applications share certain commonalities. The design and development of core power supply functions that can be used in most future space equipment will have both a commercial and a technical impact on reduction of development time, risks, and costs. Project objectives The project is divided into two main tasks: 1) Design and development of a low power instrument power supply and 2) Design and development of a high power solar array power supply using the maximum power point tracking principle to extract maximum available power from a satellite solar array. Activities 1. Low power instrument power supply: Development is concentrating on the design of a core power supply function that can be used as an off the shelf product in various space programmes. Within the framework of the project, new package technologies, especially for certain parts of the core function, are being studied. The main tasks are design, development, prototype manufacturing, and testing. 2. High Power SA Regulator: A high power SA regulator (Array Power Regulator, APR) will serve as a regulator between the solar array section and the main power bus/battery. The main design targets are very high efficiency, low mass, and modularity. The solar array will consist of many solar array sections in parallel. To ensure reliability and modularity, each solar array section will be interfaced with its own APR. The APR may comprise one DCDC converter or one high power and one lower power on top of it. Design and development comprise a study of the most efficient solution, prototyping, and testing of the selected solution. A block diagram of one module is shown in Figure 1 on page 17. Project duration: From 1 st August 2005 to 1 st August 2006. The project is still going on. Commercial impact The commercial impact has been targeted to improve Patria s readiness to comply with the requirements of future space programmes. 16

Figure 1. A block diagram of one module. Project participants The following will co-operate in the project: Patria Systems Oy Tampere University of Technology (projected DCDC converter dynamic modelling ) Technical Research Centre of Finland (projected vibration tests for new packages). Contact information Patria Systems Oy www.patria.fi Commercial matters Kimmo Myllyoja, director, space Tel. +358 20 469 2325 kimmo.myllyoja@patria.fi Technical matters Juhani Simola, project manager juhani.simola@patria.fi Tel. +358 20 469 2333 Improvement of impulse radar (ImRa) Background ImRa is high-resolution radar for short-range detection of subsurface objects and interfaces. It can be used for example to indicate rot or decay in living trees, the thickness of snow and ice, or the location of moisture faults, joists, pipes, cables etc. in a concrete, wooden or other structure. The radar can be used as a hand-held instrument in the field; it is portable and powered by rechargeable batteries. Project objectives Development of ImRa was completed in 1995. Since then there have been great advances in electronics, battery, and computer technology. The project goal was to update the radar construction and to make it easier to interpret the radar images. Activities The radar electronics were redesigned using surface-mount components and programmable logic 17

devices. The radar computer and display were replaced with modern versions. This all resulted in reduced size, weight, and power consumption. The old DOS-operating system was replaced with Linux and this enabled use of all modern memory and network options. Different digital image enhancement methods were also studied, but this still requires further development. Project duration: from 7 th January 2002 to 5 th April 2004. Results The impulse radar hardware and software were updated. The new version is smaller, lighter and more user-friendly. Project participants Helsinki University of Technology (TKK), Laboratory of Space Technology As a subcontractor, TKK carried out the study of different image enhancement methods. Contact Information Insinööritoimisto Toikka Oy Martti Toikka, managing director Tel. +358 0 813 5929 toikka@co.inet.fi www.toikkaoy.com Hyperspectral Imaging Technology Background SPECIM develops and manufactures hyperspectral imaging instruments and systems for industrial, life science, and remote sensing applications and is one of the leading companies in this area. Project Objectives The project aimed at developing advanced imaging spectrograph and camera technology in both the VNIR (400 1000 nm) and the SWIR (950 2400 nm) range. The target was to be able to meet the user requirements for higher resolution and higher quality data produced by small, economic airborne hyperspectral imagers. Activities A new optical design for high performance imaging spectrograph was completed and a spectrograph prototype in both the VNIR and the SWIR range was implemented and tested. Several VNIR cameras were evaluated and the most suitable was selected for hyperspectral imaging. A high performance SWIR camera was developed by employing new European made MCT detector technology. The spectrographs and cameras, together with a Windows-based real time image acquisition solution, were integrated with hyperspectral imager prototypes that were thoroughly characterized in the laboratory and during a flight mission. Project duration: from 1 st February 2002 to 31 st March 2003. Results Advanced, characterized technology for making high performance, compact imaging spectrographs, and high speed cameras and image acquisition solutions for hyperspectral applications in remote sensing and industry. Commercial impact SPECIM has exploited the results in three new hyperspectral imaging systems that it has developed for airborne remote sensing. AISA+ and AISA Eagle operate in the VNIR range and AISA Hawk in the SWIR range. These new systems have proven very competitive and successful in the remote sensing market. SPECIM has already sold nearly 30 of these systems world-wide and achieved a leading market position as a manufacturer of compact, high-performance, economical hyperspectral imaging systems. Also, the hyperspectral cameras for industrial and scientific applications developed by SPECIM on the basis of the new technology have had an excellent response in the market. By the end of 2005, application of the results in new products had increased SPECIM s turnover by 75%. 18

It is based on established methodology, which allows the user to develop reusable, maintainable parallel computing solutions with a mainstream language (C++) based on distributed native objects. It has a short learning curve and it is not necessary to learn new languages, paradigms, or APIs. Figure 1. An AISA Eagle hyperspectral sensor (left) acquires VNIR hyperspectral images with 1000 pixel resolution. A SWIR hyperspectral camera for industrial process analytical applications (right). ClusterMaster supports different run modes for developing and executing applications both in real clusters and in a single workstation (e.g. in a laptop). Hence, it is easy to implement software with it. It is also easy to use, because no complicated installations or external launchers are necessary for running parallel applications. Many existing real-world applications are inherently object-oriented and offer themselves to parallelization via distributed computational objects. High-level, problem-oriented solutions are available in the form of extensible, scalable, high -performance class libraries. Project participants Project duration: from 1 st 31 st August 2005. April 2004 to SPECIM carried out the project in collaboration with several of its network partners. These included R&D partners in Finland, Sweden and the USA for optical and optomechanical design, and sub-suppliers in Finland, France and Bulgaria. Contact information Contact information PIEneering Oy Jan Heikkilä, managing director jani@pieneering.fi Mobile +359 40 5098 690 http://www.pieneering.com/ SPECIM, Spectral Imaging Ltd. Timo Hyvärinen, managing director Tel. +358 8 5514495 timo.hyvarinen@specim.fi www.specim.fi ClusterMaster The ClusterMaster programming kit provides a solid base for developing new high performance computing applications or migrating legacy applications into clusters. ClusterMaster is a unique cluster programming package for developing parallel applications reliably and cost-efficiently. Figure 1. ClusterMaster run-time implements core functionality. Higher-level class libraries provide solutions for typical parallel computing scenarios. Applications can leverage both levels of abstraction. 19

A Service for defining forest patterns Background Keyfor Oy has worked with aerial-photo-based forest inventory and stand-fixing. For this, KeyFor has used an application originally developed by Arbonaut Oy. Its main customers have been Finnish and North American forest companies. Project objectives The project has two main objectives: 1. As the main focus of business has broadened, one of the objective is now to create a stand- fixing application that is better suited for various types of source data (maps and aerial photos). 2. The other is to development an even more efficient set of tools and processes for stand-fixing to speed up the whole workflow. Activities What has been accomplished: 1. Application requirements and specification. 2. Development platform evaluation to meet the following criteria: Suitability for handling of large amounts of source map data, compatibility with other customer systems, possibility to implement interfaces with other systems. 3. Research on current processes of Finnish forest companies. The stand-fixing application is under development. The initial launch is planned for summer 2006 and production usage shortly after that. The application is being developed on top of the MapInfo platform with MapBasic development language. Project duration: from April 2005 and is scheduled to end in June 2006. Results The development platform has been selected on the basis of research on the current processes of Finnish forest companies and the potential for meeting the criteria mentioned in the Activities section. Development is still under way. Commercial impacts The stand-fixing process application is/will be conform more closely to customers systems. Project participants Keyfor Oy and Arbonaut Oy Contact information Keypro/Keyfor Oy Ari Rummukainen, managing director Tel. +358 13 226 438 ari.rummukainen@keypro.fi www.keypro.fi Automatic tie points for digital aerial images Description of the project Stora Enso and Technical Research Centre of Finland (VTT) have developed a digital aerial imaging system called EnsoMosaic. The system is a complete set of software and hardware for small and medium format aerial imaging and image processing. The image processing software was based on semiautomatic search and location of tie points for image rectification. The objective of the project was to automate the tie point search for more efficient image processing that could handle very large blocks of images originating from various sensors. New algorithms were developed to link images with each other within stereo coverage and to automatically search for, evaluate and establish tie points between image pairs. The points located could be re-evaluated in a second search iteration and automatically transferred to all images within the common overlap area. Aircraft attitude information was included in the process to increase rectification speed. Project duration: from 1 st February 2001 to 31 st December 2002 Results The new tie point algorithm was successfully implemented. It reduced tie point search time by 20

Figure 1. EnsoMOSAIC user interface. about 80% and overall image processing time by about 60%, thereby enabling nearly real time orthorectification of digital images. Commercial impact The development project increased sales of the EnsoMosaic system during and especially immediately after the project. Existing EnsoMosaic users updated their software with automatic algorithms, and new EnsoMosaic operators appeared in the aerial imaging market. The commercial and customer details are confidential. Project participants The project was led by Stora Enso. VTT was subcontracted to provide the research know-how and programming skills for algorithm development. Contact information Stora Enso Wood Supply Janne Sarkeala, manager, remote sensing Tel. +358 20 4624953 janne.sarkeala@storaenso.com www.storaenso.com/ensomosaic Development of a Pseudolite System Background Space Systems Finland (SSF) is a privately held company owned by its key employees and the Finnish National Fund for Research and Development (Sitra). SSF s main products are realtime software solutions for satellites and other critical applications, as well as advanced radio navigation infrastructure. SSF is one of Europe s leading providers of space applications software. Since its formation in 1988, SSF has been providing high reliability software solutions for many of Europe s most ambitious space missions. Project objectives The goal of the project is to develop a pseudolitebased navigation system that extends the usage of satellite-based positioning methods into environments where the GNSS signal is either weak or not at all available. The target is to develop a reliable system with high accuracy for demanding applications. Furthermore, the project contains a 21

study on how the system can be applied as a local component within Galileo and what the related applications could be. One important driver for the project is Galileo and the way the Galileo mission describes pseudolites (pseudo-satellites) as an important technique for providing local augmentation. Project activities The project activities are divided into the following three main tasks: a. Research and development of algorithms for robust pseudolite navigation b. Research on Galileo pseudolites for military and civil use c. Development of the current pseudolite system to the level needed for industry applications. Project duration: from 1 st April 2004 to 31 st May 2007 Results SSF has built a navigation system consisting of synchronized pseudolites. The system design was made with the aim of supporting usage of the pseudolite system as a local service with minimal user intervention and minimal modifications to existing receiver hardware and software. The developed system can be used as a stand-alone navigation system, i.e. with no connection to GPS. The system can also be used as a local component of GPS and future Galileo. The deployed system can increase both the availability and accuracy of the ordinary navigation solution. In an indoor environment, which is generally considered very difficult due to heavy multipath, the system has shown that sub-decimeter accuracy can be achieved. By utilizing the enhanced pseudolite system, navigation algorithms developed within the project show that carrier ambiguity resolution, and hence extraordinary accuracy, is possible in real-time in these demanding environments. During the scope of the project the system has also been deployed in various outdoor environments. Several customer cases, where currently existing GPS solutions are not sufficient, have been identified. The problems experienced by these customers are mostly related to out-takes in the navigation solution or to accuracy losses due to the difficult environments (e.g. building and constructions obstructing the direct line of sight to the GPS satellites). Contact information Space Systems Finland Oy Mårten Ström, project manager Tel. +358 9 61328659 marten.strom@ssf.fi www.ssf.fi Modular extensions for ESAComp software Background Componeering Inc. develops and markets ESA- Comp software for the design and analysis of composite structures. Development of ESAComp began in 1992 with support from the European Space Agency. Since 2000, Componeering has been commercialising the software product. Since the needs of users of ESAComp software differ widely, it was decided that it should be possible to include and license add-on modules. Two add-on modules were identified as good test cases: a cylindrical shell analysis and modelling of multiaxial reinforcements. The former is of interest to the aerospace industry (e.g. launchers) and various other fields (e.g. pressure vessels). The latter is particularly of interest to marine and wind energy industries, thus addressing the technology transfer aspect. Project duration: from1 st September 2002 to 30 th June 2005 Results The capability to include add-on modules in ESAComp software was realised successfully. This has improved the prospects for creating profitable business with ESAComp. 22

The cylindrical shell module was completed to prototype level. Background work was conducted for the multiaxial reinforcement module. Both modules will be released in 2006. Although these add-on modules have not been completed, the licensing of other add-on modules has already begun. A limited market study was conducted to support the future marketing of the new add-on modules. The feedback from current users of ESAComp and indicates that interest in the planned modules does exist. Project participants A theoretical background study for the cylindrical shell analysis was conducted at the Helsinki University of Technology, Laboratory of Lightweight Structures in the form of a master s thesis. Componeering also started collaboration with CSC, the Finnish IT centre for science. Simulation technology developed by CSC is being integrated with ESAComp to realise part of the planned new capabilities. Contact information Componeering Inc. Markku Palanterä, general manager Tel. +358 9 4342 1550 markku.palantera@componeering.com www.componeering.com FDIR Applications Background and project objectives The objective of the project was to analyse industrial markets for FDIR (Failure Detection, Isolation and Recovery) technology, which was developed within ESA programmes. The technology is used in increasing the reliability of a control-system-like satellite computer system. Activities The main activity was to determine potential industrial customers in Finland that would need FDIR technology to increase the reliability of their products. The work consisted of detailed analysis of customers and their products in telecommunications, machinery, process engineering and forestry. Project duration: from 1 st August 2002 to 23 rd December 2003. Results Interesting industry applications were discovered in several different domains, for example machine automation of diesel engines, elevators and windmills. Needs were also discovered in the process industry. During the study it was obvious that FDIR technology would be tightly embedded in the customer system and would thus need significant product development efforts and in-depth knowledge of the target system. Contact information Space Systems Finland Oy Sami Laitinen, project manager Tel. +358 9 613 28 623 sami.laitinen@ssf.fi http://www.ssf.fi Business from Millimeter Waves (MILLI) Background Elektrobit Microwave (the former Ylinen Electronics) has been deeply involved in millimeter wave (radio frequencies between 30 and 300 GHz) technology for the past three decades. Most of the mm-wave projects have been related to space (e.g. Planck and Odin) or radar applications (e.g. traffic counter). The knowledge gained in these projects is being used in the MILLI project to generate new mm-wave business for Elektrobit Microwave. 23

Project objectives A Doppler radar platform is being designed in the MILLI project. The goal is to generate a general use Doppler module that can be used with minimal customization in many kinds of applications. The main applications will be in the field of industrial sensors and the emphasis is on improving the design flow. The aim is to get away from trialand-error methodology, which is widely used in mm-wave design. Component models for Gunn and Schottky diodes are being developed together with MilliLab (mm-wave research organization) and Radio Laboratory of TKK (Helsinki University of Technology). When used in computer simulations, the models allow simulation of all the blocks in either a circuit or 3D electromagnetic simulator. So far, the simulation results and measured test structures have conformed closely with each other. Design flow based on systematic simulation is essential for cost-effective customization of the Doppler module into different applications. Commercial impact Mm-wave Doppler radar has many advantages compared with its lower-frequency counterparts. Commercial solutions are rare because of the difficult technology area and expensive products. Elektrobit Microwave plans to offer Doppler radar products to several customers. All these products are based on the same platform and therefore the price can be set at an acceptable level, which can be maintained even if only a small quantity of products is ordered. Project duration: from 1 st July 2005 to 30 th September 2006 Contact information Elektrobit Microwave Oy Timo Hakala, project manager Tel. +358 40 3443685 timo.hakala@elektrobit.com www.elektrobit.com Asynchronous Test Access Port Background The Asynchronous Test Access Port project started from a discovery that enabled flexible test access for space electronics in satellites and space probes. This new invention, which is patented by Patria, is based on the well-known and commonly used IEEE-1149.1 boundary-scan testing technology. A market analysis showed that it might also have good potential in commercial on-theground market segments. Project objectives The primary project objective in the AVALI program was to make a market analysis and a business plan for commercialisation of the test access system invention. Commercial product specifications, other issues related to commercialisation, and market strategies were to be part of the business planning. Activities The project started with an analysis of the commercial boundary-scan equipment manufacturers, suppliers, and other technology providers that work in the boundary-scan testing sector. Development of boundary-scan technology is not part of Patria s core business areas, and therefore a partnership with at least one of the boundaryscan experts was necessary. After a study of the competitiveness of all the major parties in the business area, the Dutch JTAG Technologies B.V. was chosen to be Patria s strategic business and technology partner. After definition of the business model, a technology transfer was made from Patria to JTAG Technologies, which then developed a product based on the information received. The public launch of the first commercial product has been made and the technological information presented has been well received by the press and international professional audiences at several fairs and conferences concerning testing technologies. Project duration: from 19 th June 2002 to 31 st March 2004 24

Results TapCommunicator TM, the first commercial product developed by JTAG Technologies, has been introduced to the markets. The product contains Patria-owned TapSpacer TM technology that realises an asynchronous test access port. Other products, including embedded products in IP form, are being marketed and further developed. A subsequent utility model and patent applications based on further inventions that relate to the original idea to various degrees have been made. Commercial impact The project achieved its commercial objectives. There is still a vast market potential in numerous applications from hand held consumer electronics to very large, high-reliability networked systems. Project participants JTAG Technologies worked with Patria as a strategic technology partner. The company s personnel are members of numerous testing technology development and standardisation organisations and forums. Contact information Patria Systems Oy Kimmo Myllyoja, director, space Tel. +358 40 869 2325 kimmo.myllyoja@patria.fi www.patria.fi Figure 1. The first commercial product: TapCommunicatorTM by JTAG Technologies. 3.4 Research projects Environment monitoring using Earth Observation (ENVIMON) Background Environmental monitoring is an essential part of the operations of various private and public organizations such as forestry companies and civil rescue administrations. They have a basic need to gather relevant information on the state of the environment and process it so that it can be used in their decision-making processes. Since the environmental information should be as accurate, up-to-date, and inexpensive as possible, earth observation and other remote sensing data are most appropriate for the purpose. Hence, aerial and satellite images and sensor networks are constantly providing better and cheaper data that can be used in environmental monitoring. Accordingly, there has been a growing interest in recent years in environmental monitoring systems that efficiently utilize earth observation data. A particular need is for systems that implement the entire earth observation value adding chain. That is, systems that are able to master input data acquisition and management, pre-processing and analysis of the data into useful products and the delivery of the results to the user(s). Objective The mission of the Envimon project is to make the environmental monitoring practiced by the participating organizations in their operations more efficient. Accordingly, the project s main objective is to build systems to meet the environmental monitoring needs of six applications: natural disaster monitoring, forestry, maritime, traffic monitoring, repository site monitoring, and season monitoring. More concretely, the aim is to develop methods and tools for the diverse applications, including the respective software and hardware. In order to achieve the main objective a common software framework the EOFrame will be designed and implemented. The idea is that it will function as an open and flexible platform that enables fast and easy applications development with minimum effort. 25

Activities Six different applications are being studied in the project. Each of them involves two major parts: application development and prototype system implementation. The former involves developing novel and improving existing methods and techniques for the application, including their implementation in practice. The latter involves employing the resulting software/hardware together with the EOFrame platform to construct prototype environment monitoring systems that may be utilized in the applications. Depending on the maturity of the applications, the resource allocation between the application development and the system implementation varies. Hence, for some applications the majority of the required methods already exist and the work therefore focuses more on the system implementation. On the other hand, other applications require most of the resources for the method/technique development, and the implemented system will be quite plain. The application prototype systems are being built on the common EOFrame software framework, which will be designed and implemented in the project. It will, to a certain extent, support implementing all the elements of environment monitoring systems. That is, the framework enables employing diverse data sources, pre-processing, analyzing and preparing the data into proper products, and delivering them at the right time, in the right form and via the appropriate channel to different end users that use diverse terminal equipment. The general environmental monitoring concept adopted in the project is illustrated in Figure 1. Project duration: from 1 st January 2004 to 31 st April 2006. Project volume: 1,291 k. Results The general results of the project are the application prototype systems for the six applications and the EOFrame software for environment monitoring system development. The natural disaster monitoring system utilizes analysis methods and procedures to provide timely information on a disaster in different phases: mitigation, warning, response, and recovery. The prototype system focuses on detection of floods and extracting their impact areas using Synthetic Aperture Radar (SAR) images before and after disasters. The end product, image data with damaged area information, will be published through the Internet and delivered to multi-level users, from analysts and media to civil rescue, on appropriate devices such as computers and PDAs. The forestry system includes a new spectral bank concept and further developed methods for forest variable estimation and stem volume estimation. The spectral bank is a new concept that is based on storing and utilizing models (including a set of spectral and contextual features) dedicated to a certain landscape or forest type in a databank. The idea is to select and apply the best matching model for a newly acquired satellite image. The focus of estimation is on improved methods in terms of marginal utility of the system output. The maritime system focuses on the near-realtime needs of users at sea. The data sources are satellite data, weather forecast model data, observations and compiled data in chart format (ice charts). The novelty value of the system relates especially to methods and tools to provide the ships with MODIS-based information. The end products are visualized using a dedicated client, which involves tailored web pages and simple browsers. Thus, the system emphasizes the delivery part of the chain. The traffic monitoring system is a cost-effective alternative for traditional manual traffic census. Using airborne and fixed-platform derived image sequences and sophisticated image analysis and change detection methods, the traffic volume and its components (cars, buses, trucks etc.) are extracted from consecutive images based on the movements. The system implements a semi-au- 26

Figure 1. ENVIMON environmental monitoring concept. tomatic process for analyzing the images and outputs temporal and spatial traffic data for further manipulation and delivery. The repository site monitoring system is used for monitoring a geological repository site of spent nuclear fuel. The system is focused on detecting activities concerning building or road construction near the site. It is also meant to monitor quarrying activities and their environmental impacts at the repository site and in the surroundings. To allow all-weather and all-season capability, the system utilizes both optical and SAR satellite data. The main analysis methods are for interactive change detection between images that have been acquired at different times and between a map and imagery. The season monitoring system is used for monitoring the state of the nature. The system focuses on the tourist regions of Finland and the phenomena that are relevant to tourism, especially the progress of fall coloring. MODIS data together with other relevant data like fixed camera images are utilized in the monitoring. Especially, the system includes dedicated CMOS imagers with false-color mapping (including near-infrared channel) and increased dynamics, which enables accurate monitoring of changes appearing in vegetation. The system enables provision of a prototype service for tourists (as well as the tourist industry and authorities), which provides them with information about the conditions in the environment that will affect their traveling. The EOFrame software supports building of environment monitoring systems and it utilizes service-oriented architecture (SOA). The core of the EOFrame consists of workflow management. A workflow description (written in a workflow language) represents the data process used in an application and the workflow controller manages execution of the workflow. The workflow tasks and facilities are implemented as web services. The framework includes data storage, which is responsible for managing all the data and meta- 27

data used in the processing. The delivery of the results is also automated in the framework. It introduces the subscriber profile, which enables tailored delivery depending on the end-users needs, i.e., location, status, facilities, etc. The EOFrame utilizes standard interfaces and existing packages, e.g., XML based workflow languages, workflow engines, workflow editors, web services, web servers and java (J2EE). Thus, it is also able to exploit existing in-house or 3rd party software modules through these standard interfaces. Utilization plans The versions of the environmental monitoring systems or the EOFrame that are implemented in the project are not final products that can be directly utilized operationally. Rather they are prototypes that can be developed into real products in the future. With respect to deployment, the project will provide a variety of environment monitoring technologies and systems that can be exploited in separate company-driven product development and system integration projects that will be launched after Envimon. The project also includes disseminating the emerged knowledge and increasing co-operation and networking of the companies and public organizations operating in the field of environment monitoring. Project participants The research in the project will be carried out by VTT. Other project partners are Stora Enso Oyj, the National Research Institute for Earth Science and Disaster Prevention (NIED, Japan), the Shipping Enterprise Finstaship, the Finnish Maritime Administration, the City of Helsinki, the Helsinki Metropolitan Area Council (YTV), the Radiation and Nuclear Safety Authority of Finland (STUK), and Savcor Indufor Oy. Contact information VTT Jussi Ahola, research scientist Tel. +358 20 722 4506 jussi.ahola@vtt.fi http://www.vtt.fi/space/envimon Advanced Techniques for Mobile Positioning (MOT) Background Accurate positioning of mobile terminals is becoming an important functionality in wireless communications. There are two reasons for this. One is the requirement for emergency call positioning imposed by regulatory authorities and the other is the rapidly increasing commercial interest in various location-based services. Project objectives An important target of the project was to help the participating companies to utilize the European Galileo satellite-based positioning system and the modernized GPS system in their products. The general objective was to develop technologies for mobile terminal positioning. Improvements in positioning performance, accuracy, availability, and energy efficiency were the main goals in the work. Another target was also to help the participating operators to optimize their networks for location-based services and to utilize mobile location information in network planning and management. Activities The work included the development of link-level signal processing algorithms and their hardware implementations for satellite based positioning systems, as well as higher-level computational methods for location calculations combining location data from various sources. The project also developed cellular network design methodologies with a focus on positioning aspects. Project duration: 1 st February 2003 to 28 th February 2006. Project volume: 2.1 M Results Acquisition and tracking algorithms for Galileo and modernized GPS signals and an improved theoretical understanding of the Galileo signal structure and generalization of the BOC modulation principle utilized in the project. 28

Measurement-based results for radio propagation modelling in the satellite-to-indoor and pseudolite cases. Generic implementation structures for the algorithms developed and prototype implementation for a selected code acquisition method. I Improved knowledge about different Kalman filter configurations in navigation calculations. Receiver autonomous integrity monitoring (RAIM) and fault detection and exclusion (FDE) procedures for enhancing navigation system performance and avoiding erroneous navigation solutions. Navigation simulator software. Topology planning criteria for the UMTS network taking into account location technology. Publications: 10 international scientific journal articles, 40 international conference papers, 9 M.Sc. theses, and 3 doctoral theses where a significant part of the work was done under the MOT project. Utilization plans In the different areas of activities the project has produced general knowledge, new algorithms, and new implementations structures for the algorithms used, as well as comparisons of alternative solutions in terms of performance and implementation complexity. These results can be utilized by the industrial partners in future product and service development. Project participants Research partner Tampere University of Technology, Institute of Communications Engineering and Institute of Digital and Computer Systems Industrial partners Elisa Oyj European Communications Engineering Oy Fastrax Oy Nemo Technologies Oy Space Systems Finland Oy u-nav Microelectronics Finland Oy Figure 1. Two position solutions in urban environment using stand-alone GPS compared against DGPS reference. LS: basic approach, WLS: solution using RAIM and FDE procedures. 29

International co-operation University of Calgary, Canada Contact information Tampere University of Technology, Institute of Communications Engineering Markku Renfors, professor Tel. +358 3 3115 3937 markku.renfors@tut.fi www.cs.tut.fi/tlt Improving the EGNOS Service in Finland with Internet Radio Technology (IESIR) Background The first European GNSS system, EGNOS (European Geostationary Navigation Overlay Service), provides a positioning accuracy of 1 2 meters for the horizontal components and 2 3 meters for the vertical component at a level of 95%. However, a driving test carried out in 2003 in Finland along 6100 kilometres of highway and main roads indicated that the EGNOS GEO satellites are visible on only 51% of the road segments because of the low elevation angles to the GEO satellites. Hence, the availability of the EGNOS signal in space is rather low in Finland, especially for land applications such as vehicle navigation. Project objectives The objective of the project is to prototype a service that overcomes the limitations of the low elevation angles to the EGNOS GEO satellites. The objective was achieved by converting the EGNOS signal to DGPS RTCM signals and broadcasting the converted RTCM signals over the wireless Internet with the Internet radio technology. The EGNOS service will then be available without the visibilities to EGNOS GEO satellites and can be accessed via an old DGPS receiver. Project activities The project includes three major phases: 1) prototyping of the service, including the implementations of an EGNOS to DGPS converter, an Internet broadcaster, and mobile client applications in various mobile platforms including Nokia mobile phone, Microsoft smart phone, and Pocket PC, 2) integration of the system with commercial GIS mapping software, and 3) field tests. Project duration: from October 2004 to December 2005 Project volume: 133 k Results Static tests carried out at known locations have demonstrated that the solution prototyped in this project has a positioning accuracy similar to that of the standard EGNOS solution (1 2 meters for the horizontal components and 2 3 meters for the height component). The positioning availability, however, was improved to 93.5% according to the driving test carried out between Helsinki and Hanko. Hence, the solution prototyped in this project is a technology for applying the EGNOS service in Finland without the limitations of lowelevation angles to the EGNOS GEO satellites. Two publications were released and one presentation was made at an international conference related to the project. One MSc. thesis based on the research findings of the project will be completed in spring 2006. Utilization plans FGI has arranged access rights for the following companies and institutions to test the system: Nokia Oy, Space System Finland Oy, 3D-System Oy, NavaData Oy, UNavi-Micro Oy, the Technical Research Centre of Finland (VTT) and the Helsinki University of Technology. FGI will continue the operation of the prototype system and support non-commercial tests of the technology. Project participants Finnish Geodetic Institute (coordinator), Nokia Oy, Space Systems Finland, 3D Systems Oy. Contact information Finnish Geodetic Institute Ruizhi Chen, head of department of navigation and positioning Tel. +358 9 2955 5318 ruizhi.chen@fgi.fi www.fgi.fi 30

Development of the RADEFirradiation station Background At the end of the cold war, the RadHard market for space components collapsed and during the 90s their fabrication came to a virtual halt. The need for radiation testing of commercial-off-the- shelf components became critical in space projects. Project objectives The goal of the project was to develop more penetrating ion cocktails and to make a new user interface for the radiation effects facility RADEF, which is located in the accelerator laboratory of the University of Jyväskylä. The aim was to attract ESA and European space companies and organisations to Jyväskylä. Activities Tests with a TWTA transmitter (borrowed from ANL, USA) were performed and the results introduced in international conferences. A new user interface and data acquisition system were constructed and visits were made to a few leading accelerator laboratories that were involved in irradiation tests in the USA and Europe. Project duration: from September 2002 to February 2004. Project volume: 117 k Results The main result was the ESA/ESTEC contract - Utilisation of the High Energy Heavy Ion Test Facility for Component Radiation Studies. Thanks to this contract, RADEF became one of the European irradiation test laboratories coordinated by ESA (see figure). Utilisation plans The plan is to commercialise the testing activity into comprehensive test services. Project participants There was collaboration with ESA, CNES, ONERA, EADS, Saab-Ericsson Space, the Swedish Space Corporation and the network of the RADECS association during the project. Contact information University of Jyväskylä Ari Virtanen, senior researcher Mobile +358 50 541 9401 ari.virtanen@phys.jyu.fi www.phys.jyu.fi/radef Figure 1. European component irradiation fafilities. 31

Commercialization of the on-wafer test services of MilliLab Background The Millimetre-Wave Laboratory of Finland MilliLab is a joint research institute operated by VTT Technical Research Centre of Finland and the Helsinki University of Technology. One of the main functions initially planned for MilliLab has been to provide various commercial on-wafer testing services for millimetre-wave frequency components. While solid technical know-how, reputation, and facilities have been established over the years, the commercial application of the services had not been addressed in detail. Figure 1. The services provided by MilliLab includes the on-wafer characterisation of millimetre-wave integrated circuits using an automated RF probing station. Project objectives In an effort to commercialise the on-wafer test services, a project was included in the AVALI programme of the National Technology Agency (Tekes). The scope of this project can be divided into three parts: a. The accuracy and quality of measurements was to be improved with mechanical and technical improvements to the infrastructure as well as development of test methods and calculation algorithms. b. The services offered were to be more closely defined in terms of measurement uncertainties. This also includes definition of standardized operating practices and quality assurance guidelines. c. The marketing aspects were to be studied. For the marketing study, some additional funding was awarded by the TULI programme of the National Technology Agency. Activities and results Several modifications and improvements to the measurement facilities were performed during the project. These have had a major positive impact on the overall subjectivity and reliability of measurement results. Methods for the quantitative improvement of noise parameter measurement accuracy were explored, but so far, none have proved to be useful. A uniform code of conduct was formulated for the most important characterisation services; this helps in reducing systematic errors due to non-standardized measurement procedures. A more general Quality Assurance Manual was compiled that contains guidelines regarding such matters as equipment maintenance, record keeping, and the handling of devices being tested. The measurement uncertainties were investigated by performing statistical analysis of large sets of repetitive data, by comparative measurements against external facilities, and by statistical simulations. The analytical investigation of measurement uncertainties in the case of scattering and noise parameter measurements proved to be beyond the scope of this project. A marketing strategy was formulated with the help of a consultant, which provided valuable insight into the current market outlook as well as suggestions for future development of the services. Further marketing activities are to start in 2006. Project duration: from 8 th April 2003 to 7 th April 2005 Project volume: 260 k Project participants MilliLab, VTT and Pivotal Consulting Oy Contact information VTT MilliLab Arttu Luukanen, senior research scientist Tel. +358 (0)20 7224674 arttu.luukanen@vtt.fi www.vtt.fi/millilab 32

3.5 Consortium projects Dual polarization Doppler weather radar The Consortium Vaisala Oyj, the University of Helsinki (UH), and the Finnish Meteorological Institute (FMI). Background Vaisala Conventional single polarization radar detects the intensity of the backscattered signal that is used to estimate the precipitation rate and liquid water content due to the scatterers. The radial velocities of the scatterers are derived from phase differences between the transmitted and received signals. Polarimetric radar also transmits and receives variable polarizations and thus obtains information on the shapes and orientations of the scatterers. This provides new premises for classification of precipitation types and improves the quantitative measurement of precipitation. Examples of these capabilities are presented in figures 1. and 2. UH Radar meteorology has been one of the main meteorological research objects of the UH since 1965. The UH Weather Radar Laboratory has the best expertise in Finland on weather radar and is well-known worldwide. Since 2000, the UH weather radar laboratory has explored and evaluated numerous prospects for developing polarimetric research weather radar of the highest quality. These efforts led to co-operation with Vaisala and eventually to this AVALI project. Figure 1. Radar reflectivity and cross-section on 9 th August 2005 at 16:40 UTC. 33

Figure 2. Meteorological and non-meteorological scatterers identified by a polarimetry. FMI Data from weather radar networks has long been integrated for use by meteorologists by combining radar pictures in a Cartesian coordinate system. This mozaicing method is simple, but leads to data loss and other quality problems. Operational weather radar scans the atmosphere by rotating the antenna horizontally and changing the elevation angle for each full turn. Hence, the data are originally gathered in polar coordinates and should not be converted to Cartesian before being combined with data from other radar in the network. For the radar technology part of the project, the knowledge of radar installations and networking in FMI was offered to the Vaisala company. FMI s international experience of site acceptance testing of new radars was also planned for use with the new radar construction of the Vaisala company. Project duration: from 1 st May 2003 to 31 st December 2005. Project volume: The volume of research part was 553 k : 203 k for UH and 350 k for FMI. Project objectives Vaisala The purpose of this project was to build the prototype of a polarimetric weather radar employing an orthogonal dual polarization basis and intended for simultaneous transmission and reception of vertical and horizontal orthogonal polarizations. UH The responsibilities of the University of Helsinki Weather Radar Laboratory (UH) were to draw up the specifications for the prototype to participate in the planning and development of the prototype 34

to prepare and provide the premises for the prototype to participate in the installation and technical testing of the prototype to carry out meteorological testing and research using the prototype. FMI The main goal of FMI s part of the Weather Radar project was to develop a system which optimally combines the radar data at the polar coordinate level, taking quality parameters into account simultaneously. As a result, the output data will contain not only the radar data, but also a series of quality fields that can be used to estimate the effects of different error sources during measurements and post-processing. This approach provides an opportunity to fulfill the widely varying needs of customers using weather radar information. To achieve the main goal, the project also included the results of several radar-related research projects of FMI. They contributed information on different error sources and for developing correction methods. The main issues of this research were analysis of the vertical profile of reflectivity detection and classification of non-meteorological echoes 3-D distribution of the water phase in precipitation correction of attenuation in precipitation probability of hail morphological classification of precipitation type detection range of precipitation. It was also agreed that the site acceptance test (SAT) and the operational test (OT) of the new radar would be performed by FMI at the end of the project. Activities and results Vaisala The prototype Polarimetric Weather was installed at Kumpula Campus by the end of 2004 as presented in figure 3. In 2005, an extensive test and calibration period was started during which the radar participated in the Helsinki Testbed measurement campaigns. Performance of the radar prototype was observed to meet the requirements of an operational weather radar and its polarimetric capabilities proved very promising. Antenna Assembly: The antenna of the polarimetric radar is equipped with an orthomode feedhorn to allow simultaneous transmissions and receptions of horizontal and vertical polarizations. The side lobe level and cross polarization of the antenna are extremely low due to design criteria aimed for polarimetry. The pedestal of the antenna is of a yoke design, which allows location of the antenna dish close to its supporting axis. This guarantees a low weight for the antenna/pedestal structure. The pedestal and the antenna accommodate wave-guide structures and the appropriate rotary joints and slip ring connections. The rotation rate of the antenna is continuously selectable in both azimuth and elevation. All wave guide components between the transmitter and the antenna are capable of handling the full power of the transmitter. Transmitter: The transmitter of the prototype polarimetric radar is based on a full coherent klystron power amplifier fed by a stable source that can be flexibly modulated. Phases, amplitudes and lengths of the pulses transmitted as well as pulse repetition times are freely externally selectable for each pulse. 35

Figure 3. Installation of antenna assembly at Kumpula campus. A switching DC power supply and a full scale solid state cathode modulator are used to drive the klystron. In addition, a vac-ion pump controller, filament supply, and solenoid supply are required for klystron operation and protection. A full scale solid state modulator allows generation of short and long pulses by the transmitter. The pulse width and the pulse repetition frequency are continuously adjustable from pulse to pulse. UH 2003: The specifications of the Prototype were drawn up by Dr. Timo Puhakka, the director of the UH Weather Radar Laboratory. Planning of the prototype began. Preparations for the meteorological and technical tests began with measurements of the beam pattern of the UH Doppler radar. Development of a radar calibration method employing the sun began. Testing of the modules of the prototype began. The first beam-pattern measurements of the new dual polarization an- 36

tenna of the prototype were made. The results indicated that the antenna would considerably exceed the specifications. 2004: Construction of the prototype began. The prototype was first assembled and tested technically at the factory. The prototype was installed at the UH weather radar laboratory on 13 December. The radar horizon of the prototype was accurately mapped prior to installation. Software for collecting raw IQ-data and spectra from both the UH Doppler radar and the prototype was developed. This software is an essential device for research and testing. Research on the radar calibration continued (one conference publication by P. Puhakka, M. Leskinen and T. Puhakka). The UH Doppler radar was installed in a container and moved to a location 32 km north of the prototype. This UH research radar setup (Figure 4) played the key role in the radar meteorological research and testing of the prototype. 2005: On site technical tests of the prototype were performed in January after completion of the installation. Encouraging test results were confirmed by the first successful meteorological measurements performed together with the UH Doppler radar from February onwards. The measurements clearly showed the ability of the prototype to produce accurate estimates for all polarimetric parameters. The preliminary analysis of the snowstorm of March 17 th, illustrated the great potential of polarimetric measurements in mid latitude snowstorms (Figure 5). As a result, the paper Preliminary polarimetric analysis of a winter storm causing catastrophic traffic problems (by T. Puhakka), and the related paper Solar calibrations with dual polarization weather radar (by P. Puhakka) as well as a poster The UH research radar setup (by T.Puhakka) were presented at the 32 nd Conference on Radar Meteorology, organized by the American Meteorological Society, on 24 29 October in Albuquerque, New Mexico. Figure 4. The polarimetric radar prototype (on the left) measures vertical range-height cross-sections of all polarimetric parameters through the beam of the vertically pointing UH Doppler radar (on the right). 37

Figure 5. The left column of the figure shows range height cross-sections of polarimetric parameters obtained by the prototype radar on 17 th March 2005 at 06.28 UTC. The right column depicts how the precipitation particle types at all heights and ranges can be deduced. By combining this with particle fall trajectories the precipitation type on the ground can be diagnosed. The first experiences with the prototype polarimetric radar and the vertically pointing Doppler radar (the UH research radar setup) have generated numerous new research ideas. Research aspects to be initiated include advanced signal processing deriving raindrop size distributions characterizing precipitation by polarimetric measurements detecting super-cooled water and the probable regions for in flight icing of airplanes. For the new research, the site of the UH vertically pointing weather radar has been improved by in- 38

stalling there new research instruments including a complete radio sounding station, an automatic weather station, a POSS instrument, and a distrometer. FMI During the 3-year project the topological and computational aspects of the system were studied thoroughly by one researcher. At the same time the whole radar team of FMI was working on supporting projects. Several papers and posters related to these projects were presented at international radar conferences. The prototype of the system and some additional tools were programmed during the project. Operational versions of these programs will be installed during 2006 as a part of FMI s operational weather services. The SAT and OT of the prototype radar were successfully performed and reported during autumn 2005. The OT concentrated on using the new radar as a part of the operational radar network. During the 30 day testing period the radar was in practice only one in the FMI s network, so the data were collected and monitored equally. The system to combine raw radar network data for use of weather service was programmed and documented successfully and is ready to be installed in operational environment of FMI. The SAT and OT of the prototype radar of the Vaisala company were successful. Contact information Vaisala Oyj Pentti Karhunen, manager, radar technology Tel. +358 9 8949 2285 pentti.karhunen@vaisala.com www.vaisala.com University of Helsinki Timo Puhakka, director of weather laboratory Tel. +358 9 191 50870 timo.puhakka@helsinki.fi www.helsinki.fi Finnish Meteorological Institute Jarkko Koskinen, professor Tel. +358 9 1929 4174 jarkko.koskinen@fmi.fi www.fmi.fi Oil spill detection information in the Baltic Sea region (OILI) The Consortium The OILI project consortium consists of the following research institutes: the Finnish Environment Institute (SYKE), the Laboratory of Space Technology at the Helsinki University of Technology (TKK), the Finnish Institute of Marine Research (FIMR), and the company Centroid Oy. The company INPHO Technology was involved in the first phases of the project. The consortium was involved only in the OILI project, which developed BORIS (Baltic Oil Response Information System), a satellite-image system for monitoring illegal oil spills in the Baltic Sea. The project was not divided into subprojects. SYKE was responsible for the overall coordination of the project, end user requirements and systems design, implementation and integration of the final system, and development of subsystems for uncertainty assessment. TKK developed image rectification software for RADARSAT images suitable for automated processing, microwave scattering models for the sea surface, and research on detection methods for optical and microwave synthetic aperture radar (SAR) imagery. FIMR conducted research on wave modelling and observations on wave properties relevant to microwave scattering as well as research, development and integration regarding methods and corresponding software for practical oil spill detection. Centroid Oy produced demonstration web user interfaces for the system and INPHO Technology developed software for combining in situ and remote sensing data. The project was managed by the Geoinformatics and Land Use Division of SYKE. The project supervision board was chaired by Rami Ruuska from the Ministry of the Interior and it included representatives of the Data and Information Centre and the Environmental Damage Division of SYKE, FIMR, TKK, the Finnish Meteorological Institute, Tekes and the companies involved in the consortium. 39

Project duration: from1 st January 2003 to 31 st March 2006 Project volume: The volume of research part was 904 k : 579 k for SYKE, 187 k for FIMR, and 138 k for TKK. Background and objectives Being a very closed sea, the Baltic is highly susceptible to toxic substances. Oil is particularly problematic due to treatment of oil-contaminated water in ships. Unfortunately, it is common practice to wash out ship tanks and pump the water-oil mixture directly into the sea. This is of course prohibited. Prosecution, however, requires direct evidence of the incorrect procedure. The evidence is difficult to obtain, as the sea is large and the monitors few in number. In particular, the middle parts of the sea are difficult to monitor and ships take advantage of this. Remote sensing techniques are clearly in the position to combat this development. The objectives of the project were to develop manual and semi-automated remote sensing methods for detecting oil spills on the sea surface from satellite imagery, to implement a web-based system for accessing the data with common web browser, and to integrate the observations with other relevant information and drift models to assist in combating oil spills. Activities in SYKE SYKE developed methods and software for assessing the confidence level of the observations of oil spills with SAR. In addition, physical properties and changes in the oil spilled into the sea according to temperature, oil type, etc. were studied in a literature review. Automated systems for processing and net-based distribution of the images and the related data were established. This includes the capacity for automatic processing of both the SAR image data and the oil spill interpretation results provided by KSAT. Activities in TKK The Laboratory of Space Technology at the Helsinki University of Technology (TKK) developed sea surface scattering models for microwave radars and rectification software for SAR images. They also developed preliminary algorithms for slick detection from the SAR images and optical measures for detecting oil among ice. Activities in FIMR Digital video camera data collected by the R/V Aranda were analysed to quantify the dependence between the wind speed and the fraction of open sea roughened by small waves. FIMR developed algorithms and software for operational oil slick detection from SAR images. In addition to the SAR pixel values, the algorithm requires the local incidence angle and the wind speed and direction as inputs. Activities in Centroid Oy Centroid Oy developed demonstration versions of the web user interface of the BORIS system. They also participated in planning for technology transfer and commercialisation of the system. Activities in INPHO Technology Oy INPHO Technology Oy developed a prototype version of software for assimilating remote sensing and ground truth information to generic numerical environmental models. Research results BORIS (Baltic Oil Response Information System) was the main outcome of the project. BORIS is capable of distributing maps and other related GIS information relevant to oil combating tasks along with the satellite imagery and information on detected oil spills. The system uses the environmental administration s standard browser web map user interface technology with corresponding user identification systems, thus all data available in the environmental administration map services can be accessed with the same system. In addition, the BORIS system is connected to OpHespo, the operational drift modelling system of the Finnish Meteorological Institute. 40

A dedicated user interface for ArcGIS, the operator user interface, can be used to verify and modify the data to be published in BORIS. The verifications and corrections are necessary, as oil slicks are seldom easy to distinguish reliably with automatic detection methods. The automatic detection input data format and the corresponding GIS data formats store the numerical values of the features calculated for each individual slick. Thus the BORIS system constantly generates training data for new versions of detection algorithms. It is also possible to use the BORIS system to publish and redistribute to users data products and oil spill detection results acquired from an external service provider (Currently implemented: KSAT, Norway). The retrieval of satellite or other types of data from an SAR receiving station or other sources and the data processing are both automated. RADARSAT 1 data are currently processed with TKK software. In addition to satellite data, HIRLAM weather model data are acquired from the Finnish Meteorological Institute. In addition to BORIS, the OILI project developed abilities and procedures to process SAR data in SYKE. The concepts and systems developed for BORIS system and probably parts of the developed software can be used in future development of emergency response systems for the administration. Relying on the performed data analysis, the minimum wind speed for reliable oil spill detection was determined to be 4 m/s (conservative choice). During the data analysis a novel photogrammetric method was developed to extract the relative proportions from panorama images of the sea surface. This is the first time that the relationship between wave covered area and wind speed has been established in open sea conditions. Figure 1. BORIS user interface. 41

FIMR developed and implemented a method and related software to extract wave spectra from digital image sequence. The method provides a lowcost tool to determine the 2D spectra of cm-scale waves that are mainly responsible to the C-band radar backscatter. Hence, e.g., the theoretical spectra presented in the literature can be compared with the spectra occurring in varying wind conditions. The method used by SYKE for reliability assessment of SAR detection relies on wind information. The microwave SAR image detection method is essentially based on the smoothing effect of oil. As the small waves that are the main backscatterers of the microwave radiation are reduced, the oil slick area appears dark in the image. This phenomenon is disturbed in very high winds, and it does not appear if the wind is too low to generate the waves. The main contribution of FIMR was the development and implementation of a novel oil slick detection algorithm to be used within BORIS. During the algorithm development, several backscattering models were tested and one model was selected for use as part of the algorithm. The algorithm was tested using the collected database. The constructed decision rules can be modified when more oil spill/sar data pairs become available. TKK demonstrated the potential for detecting oil with airborne imaging spectrometer AISA. Lack of a suitable high resolution near real time imagery prevents operational use of optical satellite imagery. Company results and commercial impact As some of the experts left INPHO Technology Oy, development of and commercial interest in the topics specific to OILI project ceased. During the project, web map server software, which was developed and introduced by the Finnish environmental administration, was delivered to Centroid Oy. Thus they are both free to examine the practices used in developing their own web map technology and potentially to provide similar solutions to their clients in case the technology is for some reason deemed more suitable. The project includes planning for technology transfer and commercialisation. The main scenario is that if there are parties outside the Finnish environmental administration interested in managing and using information in BORIS-type oil spill monitoring systems, a dedicated project would be started. BORIS technological knowhow and a copy of the system would be transferred to Centroid Oy. In addition to their own technology, Centroid Oy already has software tools for running web map services in technology similar to those of the environmental administration. In the technology transfer process necessary modifications would possibly be made to the system to improve maintenance or the potential for local adjustments to software, for example access to different kind of web map servers. After this process, Centroid Oy would be able to provide installation, systems integration, and technical user support to future clients outside the Finnish environmental administration. Technology would be kept identical in the Finnish environmental administration and in Centroid Oy, and it would be developed jointly. Contact Information Finnish Environment Institute (SYKE) Geoinformatics and Land Use Division Timo Pyhälahti, senior expert Tel. +358 9 40300662 timo.pyhalahti@ymparisto.fi SYKE: http://www.ymparisto.fi/syke TKK: http://www.space.tkk.fi FIMR: http://www.fimr.fi Centroid Oy: http://www.centroid.fi INPHO Technology Oy: http://www.xposition.fi 42

Use of individual tree interpretation obtained from aerial images and laser scanner further development of methods and market analysis The Consortium The partners in the project were the Finnish Geodetic Institute, the Department of Remote Sensing and Photogrammetry (FGI, coordinator), the Helsinki University of Technology, the Institute of Photogrammetry and Remote Sensing (TKK), and FM-Kartta Oy (FM-K). FGI role s was the development of new methods and of change detection methods for laser scanning and transfer of technology to FM-Kartta Oy; TKK s role was analysis of the quality of laser scanning with respect to forest measurements in co-operation with FGI, and FM-Kartta Oy s role was the analysis of markets, integration of laser scanning and aerial imaging for individual tree measurements, and development of working processes. Project description and background Traditional field-based as well as photogrammetry-based forest inventory is relatively expensive and time-consuming, but relatively accurate, and thus used mainly for small-area forest inventory, i.e. standwise forest inventory in Europe. The high expenses (e.g. 20 /ha) are due to the large amount of human labor used in the data acquisition and processing. Semi-automatic and automatic techniques based on remote sensing have been extensively investigated during the last two decades to reduce costs. As a result, it has been shown that satellite images, e.g. Landsat and Spot images, can be used to derive rather reliable estimates for forest attributes in areas of more than approximately 100 hectares, i.e. those usable for national forest inventory, and that semi-automated forest inventory based on delineation of individual tree crowns (ITC) (Uuttera et al., 1998, Brandtberg and Walter, 1998, Dralle and Rudemo, 1996, Gougeon, 1997) is feasible for standwise and plotwise forest inventory. ITC-based solutions have been demonstrated with both aerial images (e.g. Gougeon and Moore 1989, Gougeon, 1997) and laser scanning point clouds (Hyyppä and Inkinen, 1999, Persson et al. 2002). Tree locations can be determined by detecting image local maxima (e.g. Geogeon and Moore 1989). In laser scanning, the aerial image is replaced by the crown DSM or the canopy height model. Provided that the filter size and image smoothing parameters are appropriate for the tree size and image resolution, the approach works relatively well with coniferous trees (Gougeon and Leckie, 2003). After finding the local maxima, the edge of the crown can be found using the processed canopy height model. The typical hypothesis for ITC based forest inventory (Hyyppä and Inkinen, 1999) is that by measuring major individual tree characteristics, such as the height of the tree, tree species, and crown diameter, it is possible to derive other valuable tree characteristics such as stem diameter, stem volume and age for the same tree and to use this information to calculate standwise forest information (mean height, dominant height, stem volume, basal area, tree species proportions). The advantage of the ITC approach is the capability to measure physical dimensions from the trees directly, and thus it requires less training material than e.g. regression-based methods (compare with e.g. Næsset, 2002). The accuracy obtained with the ITC approach depends on the structure of the forest (number of tree layers, tree species, density), the quality of images and the density of laser point clouds, and the processing technique applied. It has been shown that trees can usually be recognized only in the dominant tree layer (Persson et al. 2002, Maltamo et al. 2004b). With laser scanning, it has also been shown that multi-layered stand structures can be recognised and quantified using quantiles of laser scanner height distribution data. However, the accuracy of the results is dependent on the density of the dominant tree layer (Maltamo et al. 2005). Using ITC, Standard error at stand level typically ranges between 10 and 25% when using high-density laser point clouds 43

(Hyyppä and Inkinen 1999, Persson et al. 2002, Maltamo et al. 2004). In the case of aerial images, the corresponding figures are usually considerably higher, i.e. 35 60% (Anttila and Lehikoinen 2002, Huitu, 2005, Korpela et al. 2005). The processing of laser point clouds into forest informatics can be much more automatic than when aerial images alone are used. The costs of laser scanning are, however, significantly higher than with aerial imaging. Presently the data acquisition costs for high-density laser scanning range between 1 /ha (order size 5000 km 2 )to5 /ha (10 km 2 ). Project duration: from 1 st April 2002 to 31 st December 2004 Project volume: the volume of the research part was 255 k : 212 k for FGI and 43 k for TKK Project objectives The objective of this project was to develop a cost-effective technique for boreal forest inventory using individual tree level technology and combining use of laser scanner and aerial imagery. Research results Some examples of the project research are presented below. The effect of laser scanning parameters on DTM and tree height information The effects of the date, flight altitude, pulse mode, terrain slope, forest cover, and within plot variation on the DTM accuracy were analysed in the boreal forest zone. The boreal test site was recorded using airborne laser scanner Toposys-1 and Toposys Falcon in 1998, 2000 and 2003. Since the measurements were taken at various times during the season, i.e. May 14 th 2003 (leaf-off), June 14 th 2000 (leaf-on, low development of undergrowth), and September 2 th 1998 (leaf-on, high undergrowth), it was possible to estimate the effect of leaves and undergrowth. In 2003 flight altitudes of 400, 800 and 1500 m above ground level were used. They provided nominal pulse densities of 8 10, 4 5 and 2 3 pulses per m 2. In the boreal forest zone a random error of less than 20 cm can be obtained in most conditions for non-steep terrain. An increase in flight altitude from 400 to 1500 m increased the random error of DTM derivation by 50%. Use of first or last pulse causes a similar difference in random error. There are systematic shifts in the elevation models derived at various flight altitudes. It is expected that the beam size and sensitivity of the laser system determine this systematic behavior. Additionally, the systematic shifts between last and first pulse are significant. The difference between DTMs derived at optimum and non-optimal season conditions is typically less than 5 cm for high-density data. In stands consisting of deciduous trees, the effects are the highest. Use of a non-optimal flying season mainly causes that details of the terrain elevation can not been measured. The random error increases with increasing terrain slope. The effect of forest cover is higher when moving closer to the trunk. The results are site dependent, i.e. the accuracy obtained varies greatly as a function of site conditions. High-density airborne laser scanner data have been previously shown to provide great opportunities for individual tree detection and measurement of variables characterizing the detected trees. This study evaluated the effect of laser flight altitude on the tree height estimation at individual tree level in a boreal forest area consisting mainly of Norway spruce, Scots pine, and birch. The test area (0.5 km by 2 km) was flown over at three altitudes (400 m, 800 m and 1500 m) with a TopoSys Falcon scanner in spring 2003. Field inventory was carried out on 33 sample plots (about 30mx30m)inthetest area during summer 2001. Trees with a chest-height diameter larger than 5 cm were measured. The position, height and species of the trees were recorded; 13 plots were dominated by spruce (>50 %), 7 plots by pine, and 6 plots by deciduous trees. The rest were mixed forests. Laser point clouds in the circle of varying radius around the trees were used to extract information about spatial distribution of 44

tree crown and height without prior delineation of individual trees. Evaluations of estimation errors due to flight altitudes, including beam size and pulse density, were performed for different tree species. The results in general indicate that tree height estimation accuracy and the number of detectable trees decreases with the increase in flight height. Point density has more influence on tree height estimation than footprint size. Birch are less affected by change in flight altitude than coniferous trees. No serious effects were found on the tree height estimation in this study, at least with flight altitude changes from 400 m to 800 m. The results indicated that flight altitude could be increased to a certain level without substantial deterioration in height estimation accuracy. The costs of laser data acquisition may be reduced by increasing the flight altitude for height estimation of dominant trees. Inventory methods The objective of this study was is to improve aerial image based ITC (Individual Tree Crown) approach by laser scanning (i.e. giving the height for each crown with laser) and to compare the results obtained with high density laser scanning based ITC solutions. The study also investigated whether the inability of the ITC approach to separate tree crowns in dense forests significantly affects volume estimation. 78 carefully measured pine trees on 10 plots were used in this study. CIR image orthorectified with laser DSM was used as a major data source. Laser data were obtained from 400 m and 800 m (AGL) acquisitions providing pulse densities of approximately 10 and 5 pulses per m 2. The results of the current study demonstrated that the height of individual trees is the most important geometrical parameter for stem volume estimation that remote sensing methods can provide. Aerial image ITC approach was significantly improved by including the tree height from the laser data (for stem volume R 2 changes from 0.14 to 0.54). This kind of approach should be further studied as a potential operational forest inventory method. The results indicate that individual tree volumes can be obtained with a random error between 25 to 30% and that volume related to individual groups or segments with a random error between 34 to 40%. This paper demonstrates that even though remote sensing cannot always provide ITC solutions, the stem volume estimates do not deteriorate as a result since the existing formulas take into account widening of the tree crowns. In terms of practical forestry, it may not be so relevant whether all trees are correctly isolated with segmentation. Nevertheless, correct segmentation remains a challenging scientific task. More attention in future studies should be placed on ensuring that the size of segmented crown relates to the natural crown size and to how crown size derived by remote sensing can be converted correctly into stem diameter. It is well known that characteristics of individual trees such as tree height, biomass and crown area can be derived from airborne laser scanning (ALS) and that heights for individual trees can be obtained with an accuracy of 0.5 to 1.5 m. However, the ability to measure the growth of individual trees using ALS has not been documented. This paper reports multi-temporal laser surveys conducted in a boreal forest zone suggesting that the height growth for individual trees can be measured with an accuracy greater than 0.5 m. Methods for automatic extraction of height growth of tree crowns are presented. It is expected that similar methods are feasible for reference measurements in studies analyzing global forest changes and the carbon sink, in national forest inventories, and in describing the effects of global warming on forest growth. Other activity Thanks to joint funding from the Academy of Finland for the topic to FGI, the project resulted in 11 refereed papers in journals and international conferences and 9 other conference papers. There were 20 public presentations given on the project s results; three of them were international invited presentations. International EuroSDR/ ISPRS-supported Tree Extraction project were activated. FGI participated in organizing the Scandlaser 2003 conference. 45

There was significant academic international co-operation during the project. Also in the Tree Extraction project, which is presently supported by EuroSDR and ISPRS, there are partners from the US, Canada, the EU, and Taiwan. There has been international co-operation in writing papers, joint campaigns, numerous invitations to trusteeships etc. Company results and commercial impact A new, novel method to integrate laser-scanning data with aerial images was developed. The project supported FM-Kartta Oy in the following ways: Market analysis of the technology Technology transfer from FGI New more cost-effective ways of performing forest inventory Knowledge to optimize DTM and forest height information retrieval Submission of several patent applications. Contact information FGI Juha Hyyppä, professor Tel. +358 9 295 550 juha.hyyppa@fgi.fi www.fgi.fi FM-Kartta Oy Heikki Luukkonen, vice president Tel. +358 229 3060 heikki.luukkonen@fm-kartta.fi www.fm-kartta.fi TKK Hannu Hyyppä, senior fellow Tel. +358 9 4511 hannu.hyyppa@hut.fi www.hut.fi TKK Henrik Haggrén, professor Tel. +358 9 4511 henrik.haggren@hut.fi www.hut.fi Forest Parameter Estimation from High Resolution Remote Sensing Data The Consortium Four Finnish parties co-operated in the High- Forest project: the Technical Research Centre of Finland (VTT), the University of Helsinki (UH), Stora Enso Wood Supply (SEWS), and Fifth Element Oy (former Gisnet Solutions Finlad Oy). The HighForest project had two parallel studies: the satellite image study (VTT) and the aerial image study (UH). The Suonenjoki Research Station of the Finnish Forest Research Institute (Metla) provided the sample plot ground data (measured in summer 2001) for the system verification. Responsibilities The HighForest project had two parallel studies: the satellite image study (VTT) and the aerial image study (UH). Fifth Element s responsibility in the project was to prepare a pilot interface to demonstrate the system. The role of Stora Enso Wood Supply (SEWS) was to act as an end-user, and thus to participate in defining the scientific objectives for the study. In addition to this, SEWS specified the user requirements for the Forestime software together with VTT. VTT was responsible for the overall management of the project. Project duration: from 29 th November 2002 to 31 st January 2004. Project volume: the volume of the research part was 323 k. Background Forest owners and wood procurement organizations have a continuous need to acquire updated information on forests. The traditional standwise ground inventory is, however, expensive and its frequency is too slow. There is also a need to reduce stand sizes in order to delineate stands more flexibly when cuttings and silvicultural measures are planned, and in order to simulate future forest development more reliably. Satellite imagery with a ground resolution of twenty to thirty meters has been tested and is even used operatively in forest inventories to some extent. Although the estimates for stand characteristics are 46

reasonably good using such data, there are two major limitations that have prevented a wider operative use. The first limitation is saturation of the reflectance at higher levels of the growing stock volume. The saturation limit in Finnish conditions is at approximately 200 m 3 /ha. Another limitation is the difficulty of estimating tree species proportions since one pixel of a dimension of e.g. 30 meter often includes several tree species. The spatial resolution of remotely sensed optical imagery has increased from 30 m (Landsat Thematic mapper) to 1 m panchromatic and 4 m multispectral (Ikonos, Quickbird) in the last 20 years. At VTT s remote sensing group various methods for forestry applications using optical and radar satellite imagery have been developed within the same time span, and are in operative use as well. One goal of the HighForest study was to combine some of the existing methods into a new operative software tool utilizing the new high-resolution satellite data. The suitable application areas are all operative forest inventories as well as wood procurement mapping. Strategic planning could also utilize the information whenever information on species is required. Significant cost savings can be achieved if estimates applicable in operative forestry can be computed using imagery with a resolution of one to few meters instead of using airborne data with a resolution of tens of centimeters. By making the method for forest mapping of large forest areas fully automatic, fresh information for wood procurement purposes will be rapidly available. Relevant previous research at UH Single-tree-oriented remote sensing methods have been developed in recent years (Korpela 2000) and they have proved effective in forest attribute estimation procedures (Maltamo et al. 2002) in the forest management planning sector. High resolution aerial ortho-imageries have become cheaper than earlier and the need for replacing field work with more accurate interpretation has made them an attractive alternative. The need for detailed material is obvious where the size of trees is concerned and systems based on tree crown are being developed. However, the aerial imageries can be understood, as simulated next generation satellite imageries and increased resolution can be expected in a few years. The estimation of forest attribute parameters is based on relations between tree crown dimensions and tree diameter as well as on automatic feature extraction (Maltamo et al. 2002). Models related to required allometric tree crown dimensions are still under development; they must be completed before the methods can be applied widely. Relevant previous research at VTT Research that is associated with the spectral image interpretation for forestry purposes includes the AutoChange method for change detection and land cover classification (Häme et al. 1998) the Proba method for estimation of stand variables (Häme et al. 2000, Häme et al. 2001) the Distribution method for growing stock and conifer /deciduous tree estimation from the ongoing Modis biomass study textural measures using Independent Component Analysis (ICA) and model textures have been developed and tested with special reference to growing stock volume estimation. During the past five years, the AutoChange method for change detection and land cover classification and the Proba method for the estimation of stand variables using medium resolution data have been developed at VTT (Häme et al. 1998, Häme et al. 2001). The development of the quick distribution method for growing stock and conifer /deciduous tree estimation using medium resolution satellite data will be completed in June 2002. Project objectives The objective of the high resolution module was to develop a prototype tool for estimation of values of forest stand characteristics using satellite imagery with a ground resolution of one to a few meters. The scientific objective was to reduce the known problems of the previous satellite image aided methods, i.e. the saturation at higher biomass levels and uncertainty in tree species estimation. Also, the method should be more cost effective and faster than the methods that are using airborne data with half a meter resolution or higher. 47

The objective of the super high resolution module was to develop and demonstrate a hierarchically organized forest inventory tool set that can be used in different information requirement. Forest stands are extracted with conventional segmentation techniques and forest stand level information can be derived either by using tree crown patterns or 3D-tree measurements. Tree crown dimensions and tree size distributions were to be modeled in the first step. In the second step, the forest inventory calculation procedure was to be programmed as combination of different methods to pilot use. Project activities Activities at VTT The aim of the study was to develop a tool for the estimation of forest variables using high-resolution satellite data. The tool included modular operative software. The image analysis methodology focused on the reduction of the known problems of the previous satellite-image based methods, i.e. saturation of the estimates at higher biomass levels and uncertainty in tree species estimation. Modern contextual image analysis methods were combined with the spectral information of the imagery. In the test application the tool used images from the Ikonos satellite with ground resolutions of one and four meters. The modular software tool developed (Forestime v. 1.0) estimated the forest variables by segmenting the imagery to micro-stands, by computing stand-wise image feature vectors for the stands from the input satellite image, and by combining ground reference data with clusters from an unsupervised clustering stage. The estimates are produced as weighted sums of the input sample class probabilities. The target variables in the study were stem volume, average stem diameter, stem number, and tree species proportions (see Figure 1 and Figure 2). The first version of Forestime software was developed in the study. The guidelines in the software design were modularity, development and utilization of in-house software, and automating of the method for forest mapping of large forest areas. The modularity of the system has several benefits. It allows easy addition or replacement of e.g. some feature extraction module with another. The output of the system might as well be the land cover classification instead of parameter estimates. The system is not restricted to any specific imagery, but can fairly easily be configured to use data from a different type of instrument, or even from several instruments. The software also includes an API (Application Program Interface) for integration with other systems like GIS databases. Figure 1. Ikonos image and corresponding stem volume estimate 48

500 450 400 350 300 250 200 150 100 50 R = 0.85 0 0 100 200 300 400 500 Figure 2. Stem volume(v) estimate plotted against ground truth data. The system is a multi-tier application (see Figure 3). Forestime runs in the Windows operating system. The user interface for the System is provided by a designated GUI or external software via the System s application program interface. The server communicates with analysis tools via command line calls Forestime software was realized as object oriented design, which makes the system easily configurable and expandable. In software development Rational Rose was used for UML modeling, and JBuilder for Java implementation. The version control system was MS Visual SourceSafe. All the analysis tools were written in C. Figure 3. Forestime estimation tool. 49

Project activities at UH Different aerial photography based estimation chains were applied and tested to complete the task of forest data acquisition. We applied image measurements, interpretation and indirect estimation with models. The tree top position and crown dimensions were estimated either in 2D or 3D automatic/semiautomatic interpretation procedure and stem characteristics were estimated with allometric models. The set of independent variables of the allometric models that predict the dbh from photo-measurable quantities can differ. Calibration and validation of the models as well as photo-measurements is also possible for the control of gross systematic errors. The models for further utilization of estimated stem characteristics represent tree-level volume and taper curve functions. Forest variables are aggregated from the single-tree data. Field visits may be needed to assess the role of small undetectable trees since the photogrammetric estimates are by nature underestimates and estimates of number of small trees depends on stand structure. The system works well in well-managed forests, but has difficulties in forest stands with small clumped trees and heterogenous tree canopy structure. A model database and ArcGIS application were created to process data and calculate tree specific information using the new model chain. Project activities at SEWS The role of Stora Enso Wood Supply (SEWS) was to act as an end-user, and thus to participate in defining the scientific objectives for the study. In addition to this, SEWS specified the user requirements for the Forestime software together with VTT by defining the general use cases of the software and the product output format, and evaluated the estimation of result usefulness from the end-user point-of-view. SEWS also had an active role in specifying the Forestime software interface to SimSilva software (by Fifth Element). Project activities at Fifth Element Gisnet produced an interface to their own product SimSilva which is integrated with Forestime software for producing forest variable estimates, and for transferring the results to SiomSilva geodatabase. Gisnet took part in definition and data modeling in co-operation with VTT, and also produced the technical specification. In addition, their work included the Prototype implementation, which consisted of MIF-data import, a raster image opening and analyzing interface (SimSilva ForestTime API), data model modifications, and data import and testing. Results Research results VTT In the HighForest study VTT implemented the prototype version of a modular software tool for forest variable estimation (Forestime v. 1.0). Even if the functionality, speed, and resource consumption of the implemented software are still very limited, the operational tests have shown the feasibility of the system. The software was tested as a stand-alone system, as well as an external application program from commercial software through the API. Operation of Forestime met expectations and the objective for the software implementation was successfully achieved. The accuracies (RMSE%) of the estimated variables when compared with ground reference data were for total stem volume 37.4% (% of mean), for average stem diameter 23.4%, for stem number 87%, for pine percentage 111%, for spruce percentage 47%, and for broad-leaved tree percentage 137%. A summary of the results of forest variable estimation with Forestime v 1.0 shows that the objective of reducing the reflectance saturation effect was partly achieved, as well as the objective for accurate tree species estimation. The best results were achieved using the spectral channels and the Haralick entropy as input features. These are the recommended inputs for the use of the system in its present form. The estimation speed can still be improved, if only the spectral channels are used. In the produced forest variable estimations, the target data variance in the clusters is relatively large, which leads to averaged estimates. One main issue in future studies will be the potential for reducing the variance using different kind of 50

clustering, e.g. by taking the target variable information into the process and its effect on estimate accuracy. The future development of the system should also focus on taking full advantage of the contextual features in segmentation, and on introducing means for easy ground reference data exploration. The development of feature bank concept for the wide operational use of the system is also regarded as one key development areas of the Forestime system. One article was published from the results of project. Research Results from UH The results of this task were in the form of scientific articles and a forest information system. In an image-based measurement scheme the following are the basic tasks that need to be accomplished by the system. tree detection (~isolation of crowns, detection of tree tops) measurement of crown dimensions species recognition measurement of tree heights allometric estimation of stem sizes. The system is operated with monoscopic imagery, i.e. several views are available for each target. With single images, the system operates in the 2D image domain only and an extension into a 3D measurement scheme is possible if information from several images is used by the system. This study focuses on different data acquisition schemes in tree volume estimation and integration of these procedure with forestry GIS. Different estimation chains were applied and tested to complete the task of forest data acquisition. We also aim to bridge the gap between geographical information systems and the model chains used in data processing, parsing, and linking simulation models with datasets. Five articles were published on the results of project. Figure 4. Snapshop of ArcGIS application created during the project. 51

Company results and commercial impact Results for SEWS The main result of VTT s contribution to SEWS was the Forestime prototype program for producing forest variable estimates at stand level. The prototype program was available for SEWS test use, and it is being currently used in the ENVI- MON-project as a starting point for developing the first version of an operative tool. Results for Fifth Element The result of the project is that the VTT Forest- Time analyzing modules can be utilized together with the SimSilva forest management system. This Proof-of-Concept project supports SimSilva sales and marketing activities. Contact information VTT Heikki Astola, project manager Tel. +358 20 722 4338 heikki.astola@vtt.fi http://www.vtt.fi University of Helsinki Timo Tokola, professor Tel. +358 9 19158180 timo.tokola@helsinki.fi www.helsinki.fi Stora Enso Wood Supply Janne Sarkeala, manager, remote sensing Tel. +358 20 4624953 janne.sarkeala@storaenso.com www.ensomosaic.com Fifth Element Oy Perttu Aunola, industry manager Tel. +358 40 7743592 perttu.aunola@fifthelement.fi www.fifthelement.fi Figure 5. SimSilva user interface. 52

Quality of laser scanning, especially in urban environments (LAQU) The Consortium The Laqu team was made up of organisations from both Spain and Finland. The team consists of: Institut Cartogràfic de Catalunya (ICC), Spain, Terrasolid Oy, the Finnish Geodetic Institute (FGI) and the Helsinki University of Technology (TKK). Each team member brings particular expertise to the consortium. ICC represents a service provider for laser scanning, Terrasolid a SW company, and FGI and TKK represent research organisations. The project included a national project (Tekes), but also a Eureka project (E! 3018) together with ICC. In addition, the project included the coordination of an international building Extraction comparison of EuroSDR. The partners of the EuroSDR comparison were Finnish Geodetic Institute Stuttgart University of Applied Sciences International Institute for Geo-Information Science and Earth Observation (ITC) Helsinki University of Technology Dresden University of Technology, Institute of Photogrammetry and Remote Sensing CyberCity AG Swedish Defence Research Agency Institut Cartogràfic de Catalunya (ICC) IGN, MATIS Hamburg University of Applied Sciences Nebel & Partner GmbH C+B Technik GmbH University of Aalborg Project duration: from 1 st January 2003 to 31 st December 2004 Project volume: The volume of research part was 150 k : 50k for Tkk and 100 for FMI. Background Laser scanner survey provides a cloud of points, the x, y and z coordinates of which are known. They form a digital surface model (DSM), which includes e.g. terrain points, vegetation points, and points reflected from buildings. By processing the data and classifying the points into terrain and vegetation points, it was possible to produce various kinds of 3D surfaces representing terrain and target properties. This knowledge can be used for a large number of applications, ranging from 3D city models to characterization of individual tree canopies. The laser scanning became more famous in the mid-90s; since then the technology has been developed rapidly. Airborne laser scanning (ALS) has become an established tool for acquiring a digital terrain model (DTM) and is increasingly used for city modelling. The quality of DTM and city models derived from laser scanning is also influenced by a large number of measurement parameters. From the users point of view, the most attractive issues concerning specifications of the data acquisition are date, point density, flight altitude, and scan angle. The scan angle and flight altitude both affect the point density. Laser scanning provides a large number of dummy points requiring intelligent processing algorithms and software to do the processing. Typical project consists of tens of millions of points that cannot be handled by existing civil engineering software. Due to the fast penetration into the market, the development of commercially available software has hindered market growth. Presently, the TerraScan is assumed to be the leading global market software for ALS data processing. It is used by the majority of the users of the major laser data providers such as Optech, L/H and TopEye. TerraSolid Oy has TerraScan software, which was basically programmed by the software specialist Arttu Soininen. The global use of laser scanning is increasing at an incredible rate and reliable, user-friend software, which is based on research findings, is needed. Project objectives Research objective 1: The quality of the laser scanning is a complex problem due to a wide variety of sensors and their parameters. In the analysis of quality, the emphasis is put on the analysis of factors affecting the quality. Research objective 2: The ICC has developed an automatic procedure for adjustment of laser strips and a flight methodology. TerraSolid and 53

ICC methodologies will be compared and hopefully a new improved method could be proposed. Research results Examples from quality analysis of laser scanners in an urban environment The TIN densification method and Terrascan software were applied to derive the digital terrain model. In the Otaniemi test area, helicopter- and aeroplane-based scanning systems were used. The accuracy of the obtained triangulated digital terrain models is shown in Table 1. Table 1 shows the accuracy obtained using TopEye and Toposys flights from flight altitudes of 200, 400, 550, and 800 m. The flight altitudes of 400 m obtained from aircraft and 200 m obtained from the helicopter seem to be feasible for giving accurate digital terrain elevation information for road design purposes. These models are almost comparable to those obtained with RTK. Figure 1 shows the example accuracy analysis for TopEye (altitudes 200 and 550 m) system. The accuracy was defined for four target materials: asphalt, grass (lawn), gravel and natural (covered) terrain. The random errors obtained from 200 m are approximately 0.04 0.06 m for hard targets. For natural, modulating terrain the corresponding accuracy was 0.08 m. The verification of pure flight height is not unambiguous, because the higher the flight height the sparser the point density and the larger the footprint (beam size) on the ground. In this study in the Otaniemi area, the flying heights used were 200 and 550 m with TopEye and 400 m and 800 m with TopoSys. TopoSys flight lines were individual strips and the TopEye flight consisted of 20 strips. A laser scanner produces accurate 3D models and continuous cross-sections of terrain and auxiliary information. Demonstrations of some structures that can be extracted for road inventory are shown in Figure 3. The left image shows a parking lot and the right image shows individual trees and a building. The tree species, height and width of the trees can be determined. The low local variability of the data provides credence for the quality of laser scanning data. Table 1. Accuracy of laser DTMs. Mean height difference (systematic error) and standard deviation (random error). The accuracy was defined separately for asphalt and other surface types. Terrain type All consists of gravel, asphalt, grass and other terrain. Test Site Type of Terrain Mean height differences (m) Standard deviation (m) Pulse density (/m 2 ) Flight altitude (m) Reference Leaf on/off Laser sensor Otaniemi All 0.014 0.068 2.6 200 RTK On TopEye Otaniemi Asphalt 0.001 0.042 2.6 200 RTK On TopEye Otaniemi All -0.008 0.114 1 550 RTK On TopEye Otaniemi Asphalt -0.022 0.100 1 550 RTK On TopEye Otaniemi All 0.034 0.035 10 400 RTK On TopoSys Otaniemi Asphalt 0.002 0.042 10 400 RTK On TopoSys Otaniemi All 0.050 0.122 3 800 RTK On TopoSys Otaniemi Asphalt 0.010 0.048 3 800 RTK On TopoSys Espoonlahti Asphalt -0.031 0.026 10 400 tacheometer Off TopoSys Espoonlahti Gravel -0.020 0.016 10 400 tacheometer Off TopoSys 54

0.2 0.15 0.1 Asphalt Grass Gravel Terrain 200 m 550 m 0.05 0-0.05 z (cm) Average Std Average Std Figure 1. Effect of flight height on four target materials using TopEye laser scanner (different number of hits from Table 1). Figure 2. Demonstration of accurate cross-sections. The capability to characterize the height of man-made structures was examined using walls, fences, and traffic lights. The height of the traffic lights was obtained with a systematic error of 0.01 m and a random error of 0.15 m. Walls and fences were obtained with corresponding errors of 0.03 and 0.04 m with the TopEye 200 m data at Otaniemi. Examples from EuroSDR Building Extraction comparison This EuroSDR Building Extraction sub-study compared the performance of photogrammetric, laser scanning-based and hybrid methods on building extraction, especially on the determination of building outlines, lengths, and roof inclination. The results confirmed with experiments that laser scanning is more suitable for deriving building heights, extracting planar roof faces, and the ridges of the roofs, whereas photogrammetry and aerial images are more suitable for building outline and length determination. In building outline determination, point density, shadowing of trees and complexity of the structure were the major reasons for site wise variations of the laser scanner based results. In building length determination with laser scanning, the complexity of the buildings was the major cause for site wise variation rather than the point density. Height determination accuracy followed the laser scanning point density exactly. With high-density data in Espoonlahti, all participants were able to provide average height with accuracy better than 20 cm IQR value. Roof inclination determination was more 55

accurate when using laser data than photogrammetry, but there is still a large variation in quality due to methods and test sites (i.e. complex buildings). In general, the target plane accuracy is affected by the degree of automation. The target height accuracy seems to be almost independent of the degree of automation. CyberCity, Stuttgart and TerraScan (provided by the ICC) solutions provided the highest accuracy. There seemed to be a higher variation in the quality with other models depending on the test site or remotely sensed information. Other activities The project has the following additional impact more than 10 refereed papers in international journals and conferences 7 other papers the Lic.Sc. degree of Petri Rönnholm more than 20 oral presentations on the project coordination of EuroSDR Building Extraction comparison manuscript for the Handbook of the Laser Scanning Quality first specifications for laser scanning in Finland several invited international presentations invitations to high trusteeships at international level (EuroSDR, ISPRS) joint laser scanning campaign with TopEye AB (SWE) joint laser scanning campaign with Toposys GmbH (GER) visits by the international professors Kraus and Maas. Company results and commercial impact As an example of company results, the virtual reality improvements of the Terrascan are depicted below. Terrascan was programmed to find point groups matching a tree shape. Instead of laser scanning point clouds, an RCP (Rich Photorealistic Content) cell is placed in the model for each tree. The RCP cells consist of 300 600 photos taken of sample trees that are used to build up the virtual trees. Figure 3. Ring I visualization using TerraScan in digital terrain model and in the extraction of buildings and using TerraPhoto for combining digital aerial photos. 56

Figure 4. Clip-off of the Otaniemi fly-through animation done with ustation + Terra software using laser scanner data and simultaneously acquired aerial images. Figure 5. View of Otaniemi animation using laser scanner points in building and tree heights and texture from aerial images. The indirect impact on the whole surveying industry is significant through the use of Terrascan. More accurate DTM, for example, enables more accurate environmental analyses, better visualization, precise mass calculation, etc. The results of this research encourage surveying companies to increase the use of ALS data for mapping purposes. Examples are shown in Figures 3 5. By this means, a virtual reality of the environmental impact of a road construction process can be visualized from the defined perspective and with flythrough movies. Contact information TKK Hannu Hyyppä, senior fellow Tel. +358 9 4511 hannu.hyyppa@hut.fi, www.hut.fi Terrasolid Oy Esa Haapa-aho, managing director Tel. +358 9 637 526 info@terrasolid.fi www.terrasolid.fi Terrasolid Oy Arttu Soininen Arttu.Soininen@kolumbus.fi www.terrasolid.fi FGI Juha Hyyppä, professor Tel. +358 9 295 550 juha.hyyppa@fgi.fi www.fgi.fi 57

Statistically calibrated satellite image interpretation for land use and forest characteristics The Consortium and projects The consortium had two subprojects. the first one focusd on product development and marketing material and the second developed feasible and cost-effective methods for defined objectives. The following four partners participated in the project: WM-Data Oy (former Novosat Oy.), University of Helsinki (UH), Department of Forest Resource Management, Tmi Timo Pekkonen, National Land Survey.of Finland. Responsibilities WM-data Ltd focused on product development and marketing material. The University partner developing feasible and cost-effective methods. Tmi Timo Pekkonen was a subcontractor for software development. The National Land Survey provided digital map material and participated in the needs analysis. The project manager was prof. Timo Tokola and the project manager in WM-data was Juho Heikkilä. Project duration: from 1 st July 2003 to 30 th December 2004. Project volume: volume of resear part was 124k Background WM-Data Oy Finland has several land-use products, but more efficient methods are required to reduce the need for field-work and to develop cost-efficient products for different purposes. The products need testing with potential clients and the specific needs of potential clients are reviewed. The University of Helsinki (UH), forestry: In Finland, environmental authorities, paper companies, and teleoperators use regional forest maps on scales of 1:50,0001:100,000 for planning tasks as well as for timber procurement. A largearea forest inventory based on Landsat image interpretation is often an appropriate method for producing such maps. UH has a long tradition in developing innovative techniques for remote sensing applications and many applications have been widely used in practical forestry. Project objectives WM-Data Oy needed improvements in product development, marketing, and sales. Operative procedures for product production require quality testing. University of Helsinki: Specific issues require improvement before significant cost savings can be achieved (reliable multi-temporal image mosaic, stable stand volume models with good extrapolation characteristics, and calibration methods to utilize all available digital auxiliary data) Activities WM-Data Oy tested the procedures that were developed at UH and developed a user-interface for Erdas Imagine Software for production purposes. Detailed procedures were developed for data integration with the National Land Survey. Potential customers were reviewed and pricing mechanisms were developed. In the research sector, various atmospheric correction methods were tested to create a multitemporal image mosaic covering the target area and to apply the same interpretation procedure to the entire area. The method can be used when a large area covering the same field dataset (training data) is used for the entire area of the image mosaic. The study focuses mainly on evaluating the accuracy and functionality of stand-volume models in multi-temporal Landsat images after atmospheric correction and image calibration. Other variables of interest include height and the proportion of different tree species. The models predicting the proportion of different tree species are used to estimate the dominant tree species. Also, we compared methods for treeless peatland detection from a Landsat ETM+ satellite image. The study area was in the southern aapa mire zone in Finland. The classification methods tested included sequential maximum a posteriori (SMAP), supervised maximum likelihood (ML), and unsupervised ML with Iso-Cluster-based 58

signatures. Specific procedures were developed for calibration of final results. Two alternative methods were tested: plot-specific and area-specific. Research results Different methods were developed to prepare national level forest data from medium resolution satellite images and alternative forest data production schemes were tested. Using remote sensing to monitor large forest areas usually requires large field datasets. The need for extensive data collection can be reduced through interpretation of several images simultaneously. Various atmospheric correction methods were tested to generalize field information outside the coverage of single images. Aerosol data from MODIS were used in retrieving parameters for the 6S algorithm. In comparison with the uncorrected data, the relative RMSE values for the multi-temporal images decreased by an average of 6% with DOS, 14% with SMAC, and 15% with 6S. Figure 1. Sample product. In satellite image-based estimation and classification of forest variables in Finland, peat-lands are usually processed separately from mineral soil forests to improve the accuracy of the results. The classification methods tested included sequential maximum a posteriori (SMAP), supervised maximum likelihood (ML), and unsupervised ML with Iso-Cluster-based signatures. The existing peatland mask was superior in accuracy, but its errors could be spotted by comparing with either the SMAP or the ML results. SMAP produced more usable maps by forming large interconnected regions of the same class. During the project, a new implementation of the knn algorithm is presented for forest inventory, where the highly efficient use of programming code speeds up the process, without missing the true neighbours. The methodologies developed during the project were utilised to produce vegetation maps for planning of network for mobile telecommunication. After completing the study, the commercial partner WM-Data produced the entire coverage of Finland for TeliaSonera. Figure 2. Optimal use of field information was studied in the project. 59

Two manuscripts have been submitted to journal publications. Company results and commercial impact The company has ready-made procedures and made contact with TeliaSonera. They produced a land-use and forest map of Finland for mobile phone network planning purposes. Contact information WM-Data Oy Stefan Winqvist, project director Tel. +358 40 5140066 stefan.winqvist@wmdata.fi http://www.wmdata.fi/wmwebb University of Helsinki, Forestry Timo Tokola, professor Tel. +358 40 7002118 timo.tokola@helsinki.fi, http://www.honeybee.helsinki.fi/gis/refe/ index Interpretation methods for waveform laser scanning (Waveform) The Consortium The team consists of Terrasolid Oy, the Finnish Geodetic Institute (FGI) and the Helsinki University of Technology (TKK). Each team member brings particular expertise to the consortium. Terrasolid represents an SW company and FGI and TKK represent research organisations. The potential end-users Soil and Water Oy, Sito Oy, FM-Kartta Oy, and the Finnish Road Administration are on the steering committee of the project. Project duration: The duration of the ongoing project is from 1 st August 2004 to 31 st August 2007 Project volume: the volume of the research part was 200 k : for FGI 120k and for TKK 80k. Background The use of laser scanner intensity is limited to that of a predictor e.g. for tree species classification (Holmgren et al. 2004) or for matching aerial imagery and laser scanner data. The more effective use of intensity values is missing, partly due to a lack of calibration techniques. Full-waveform digitising lidar systems have also been developed, firstly as preparation for future satellite systems to survey earth topography and vegetation cover, and later for airborne laser scanners. Today, several manufacturers are announcing the use of airborne laser scanner capable of recording digitised waveforms. Drake et al. (2002) depict several statistical predictors that use waveform information such as lidar canopy height, height of median energy, and ground return ratio. According to Wagner et al. (2004), the use of full-waveform in airborne laser scanner offers an opportunity to classify the data based on the shape of the echo. Another important advantage is that the detection of the trigger pulses can be applied after data capturing. It is expected that the waveform of the signal will be included in most of the new laser scanners. Project objectives The objective of the project is to develop retrieval methods based on waveforms to be implemented in TerraScan SW. First results First experiences of full-waveform laser scanner data structure were gained during the 2005 research period. The Remmingtorp test area in Sweden was scanned with the TopEye-system. The area is a mixed forest, approximately 1500 long and 150 wide. Both Mk I and MK II scanners were operated during the flight and the full waveform data mode was recorded with an MK II receiver. MK II measurement data were stored in two format types, LAS-files including traditional first and last pulses (extracted from signal) and TEW-files including pulse origins, pulse direc- 60

tion and 128 intensity samples at 30 cm intervals, the full signal. As the TEW files are not yet compatible with any computer software, the work has mostly been a programming application, which would be able to read points based on a data format description sent by the Topeye company. The application was programmed and tested with the Remmigtorp data. Points were opened in MAT- LAB. However, the points were in the original measurement co-ordination system, on the wrong scale, and transformation to the Swedish RT90 75 system had to be carried out in co-operation with the Topeye company. Two approaches were considered in analyzing the full signal: classification of pulses based on signal features and new 3D point extraction from pulse. The classification exploits principal component analysis (PCA) carried out for the signal. The classes extracted were for example vegetation and ground. The other approach finds the maximum intensity of the signal and produces a new measurement point utilizing information about pulse origin location, number of intervals, and vector direction. These points are used as normal measurement points in the further process. Both approaches need further work, as only the first experience has been gained. Contact information FGI Juha Hyyppä, professor Tel. +358 9 295 550 juha.hyyppa@fgi.fi www.fgi.fi TKK Hannu Hyyppä, senior fellow Tel. +358 9 4511 hannu.hyyppa@hut.fi, www.hut.fi Terrasolid Oy Esa Haapa-aho, managing director Tel. +358 9 637 526 info@terrasolid.fi www.terrasolid.fi Terrasolid Oy Arttu Soininen Arttu.Soininen@kolumbus.fi www.terrasolid.fi Helsinki Testbed The Consortium The Consortium consists of the Finnish Meteorological Institute (FMI), Vaisala Oyj, Nokia Oyj, the Radiation and Nuclear Safety Authority Finland (STUK), the Finnish Road Administration, Road Enterprise, Helsinki Metropolitan Area Council (YTV), the Finnish Athletics Organization (SUL), and anonymous corporate partners. The Helsinki Testbed (HTB) project includes two sub-projects led by the Finnish Meteorological Institute and Vaisala. The Consortium activities are coordinated overall by FMI. In the FMI sub-project, Tekes is the main financier whereas FMI is the other significant contributing member. Except for Vaisala, the abovementioned Consortium members provide minor monetary support. FMI s tasks include coordination, planning of observation sites, design and implementation of a data warehouse, design and implementation of data distribution, and atmospheric research. The tasks of the Vaisala sub-project are design and implementation of a dense mesoscale observation network and collection of observation data and their real time delivery to the Helsinki Testbed Central Data Warehouse. An important part of the project is the development of a new mesoscale observation network architecture, utilizing platform independent software components and modern wireless M2M communication technologies. The sub-project is financed by Vaisala and Tekes and implemented by Vaisala with several subcontractors. An important contribution comes from the telecommunication base station administrator Unibase Oy, whose base station masts are being used as installation platforms for the weather transmitters. Project duration: from 1 st September 2004 and will last until 1 st September 2007. Project volume: The volume of reseach part is 460 k 61

Project description Local weather is often meteorologically referred to as a mesoscale phenomenon. These events typically cover an area ranging between 1 and 30 km and last less than three hours. They are too large to be described in one point, but too small to be readily detected with a conventional weather observation network. These types of phenomena include precipitation showers, lightning, temperature inversions, sea breeze, and fog. Although they are short-lived and localized, they can have significant socio-economic impact. In order to study them and to develop new weather and air quality services and businesses, an observation network that is spatially and temporally denser than usual is required. Testbeds provide the ideal framework for research and complete end-to-end business solution demonstrations of atmospheric information production processes. The existing observation network has been supplemented with a lot of new observation equipment in an area of approximately 150 km x 150 km around Greater Helsinki. Cellphone network masts have been used extensively as measurement sites. Hence, a tremendous amount of new meteorologically important information is obtained from the lowest 100 m thick surface layer of the atmosphere. In addition to more than hundred new automatic weather stations, new network equipment includes radio soundings, weighing rain gauges and remote sensing instruments such as laser radars, a dual-polarization radar, vertically pointing Doppler radars, and a wind profiler. Measurement campaigns have been thematically named according to phenomena or activities typical of a season, i.e. Aug 2005 Nowcasting, Nov 2005 Precipitation type, Jan- Feb 2006 Stable boundary layer, May 2006 Sea breeze, and Aug 2006 Convection. Data gathered during measurement campaigns will benefit worldwide research for years to come, as the data measured in the project will be available to anyone interested. This data will help in comparing different models for predicting atmospheric phenomena and in evaluating increased numbers of observations. Different system architectures and observation methods can be compared and the techniques most appropriate for each atmospheric phenomenon can be identified. Results from restricted geographical area can be generalized to wider national and international levels in a cost-effective way. Research results, instrument and system tests will lead to new weather services in Finland. This will benefit citizens, companies and other organizations operating in the region. A realistic experimentation environment also provides prerequisites for new export products. These can be information distribution services, complete observation networks, measurement systems, or single instruments. The aim at FMI and Vaisala is to maintain and develop the HTB concept for at least 5 to 10 years. Research results At the time of the writing, the project is ongoing and its focus is shifting away from infrastructure development towards research. The most important project result thus far can be best demonstrated as follows: An exceptionally dense mesoscale observation network has been established with a real time data web site; Fig. 1 illustrates the HTB region. For observations networks, a complete measurement process has been outlined and 24/7 supervision of the measurement process has been applied. A large mesoscale research data set is being built up. It includes unique data from a large number of communication masts at more than one level and offers excellent opportunities for atmospheric service product development. Definition of a meteorological XML-based data exchange interface, allowing a significantly improved level of modularity. This will lead to easier interfacing with third party organizations, speed up integration of future data sources, and lower the threshold in software outsourcing. There will be a skeleton model of an entirely new multi-tier service oriented meteorological network architecture, ready to be expanded to operations outside HTB. 62

Wide international publicity has been achieved within the meteorological community (e.g. WMO) as well as national publicity in the media. So far, international observational implementations in HTB have been performed or agreed with the UK and Canada. Much of the testbed ideology has been realized. Project partners representing many aspects of society have been involved. An essential part of the project has been integration and testing of network components and translation of R&D results into usable information products. Atmospheric and ICT technology integration has been demonstrated with notable cases of mobile WAP and software clients. Discussions of wider acceptance of testbed technologies have been initiated. All in all, a pseudo-operational network platform has been achieved. Company results and commercial impact The main results of the Vaisala sub-project are development of a platform independent software architecture and communications protocol for mesoscale observation networks utilizing TCP/IP. a cost-effective, COTS-based hardware solution to interface weather sensors to TCP/IP. realization of the Helsinki Testbed dense weather station network using the above mentioned SW and HW solutions. The network consists of over 100 Vaisala WXT510 Weather Transmitters, together with additional instrumentation including ceilometers, all-weather precipitation sensors, and a wind profiler. The data are collected in a Vaisala MetMan database and delivered to the HTB Central Data Warehouse using the XML-based data exchange interface developed in the project. Figure 1. Example of the real time HTB website, exhibiting precipitation intensity (dbz) from Doppler radars and surface temperature at the height of 2 m agl. 63

The project creates a technological basis for future commercial solutions in integrated and mesoscale weather observing systems, nowcasting, and precision weather applications. The application areas include public safety, traffic (road, air and sea), air quality, agriculture, energy production, and a multitude of work- and leisure-related personalized weather services. Contact information FMI Jarkko Koskinen, professor Tel. +358 9 1929 4174 jarkko.koskinen@fmi.fi FMI Jarmo Koistinen, project manager Tel. +358 9 1929 3618 jarmo.koistinen@fmi.fi Vaisala Oyj Jussi Mykkänen, director Tel. +358 9 8949 2207 jussi.mykkanen@vaisala.com Vaisala Oyj Heikki Turtiainen, project manager Tel. +358 9 8949 2261 heikki.turtiainen@vaisala.com Project www -site: http://testbed.fmi.fi GaAs Imaging X-ray Detectors Project description and background In digital x-ray imaging, especially in mammography, there is a continuous need for better detector technology to improve image quality and to reduce patient radiation doses. Planmed Oy develops, manufactures, and markets advanced mammography equipment. It is the fourth biggest manufacturer in the field. Oxford Instruments Analytical Oy has developed pixellated GaAs X-ray detector technology in its ESA TRP and GSTP projects. The technology was designed for ESA s BepiColombo mission to Mercury. OIA has developed the technology together with VTT, Modulight Oy, the Helsinki University of Technology and the Tampere University of Technology. The GaAs detectors have been processed at Modulight Oy. VTT Information Technology has bonded the pixellated detectors to readout ASIC chips. VTT is a leading developer and provider of small size (<100 µm) solder bump bonding technology. Project objectives The main target of the project was to develop x-ray detection technology for the digital mammography application of Planmed by utilizing and further developing the pixellated GaAs x-ray detector technology of OIA and bump bonding technology of VTT. The consortium The consortium consisted of 3 partners: Planmed Oy, Oxford Instruments Analytical Oy (later OIA), and VTT. The main target of the project was to utilize and further develop the pixellated GaAs x-ray detector technology that OIA has developed in its ESA projects with VTT and Modulight Oy for the mammography applications of Planmed. The consortium was led by Planmed. It stated the business related goals of the project: the volume, price levels, and technical specifications for the x-ray detectors. It was also responsible for development of the ASIC read out chip for the GaAs pixel detectors. OIA developed the GaAs detector fabrication technology, together with Modulight Oy as its subcontractor, for the mammography applications. It also studied possible other applications for its GaAs detector technology. VTT developed its solder bump bond technology to be suitable for the small pixel size required in mammography applications. Project duration: from 1 st October 2002 to 31 st December 2005 Project volume: the volume of the research part was 163 k. 64

Planmed s results and commercial impact Planmed s responsibility in this consortium project was to coordinate the project and develop the required readout circuit. Planmed was also to be the main party to utilize the results for digital mammographic imaging. A prototype of the readout chip was developed to test the various functions required. The principle of operation is fully digital, each pixel counting the absolute number of X-Ray quanta that hit on it. The chip is shown in the figure 1. A readout matrix of 3x4 pixels can be seen on the left side, next to the large black fill-in area. Separate test amplifiers were placed on the right side, but they are too small to be seen. Figure 1. The prototype of the imaging readout chip developed in the project. In figure 2 is a close up of the imaging matrix. Pixel size is 50 µm, divided in two sub-pixels. The imaging matrix digital functions were tested and found to be completely operative. In the tests, the amplifier section was found to nearly fulfil the requirements, although some fine-tuning was required to achieve optimal performance. Bias currents had to be reduced for overall power consumption and the frequency response altered slightly. After one design round the goal was to design the final readout chip. After the encouraging results the project was not continued further at this point because of variable quality of GaAs material. The design was put on hold until the material issue has been solved. The final commercial impact of this project for Planmed awaits final results. Oxford Instruments Analytical: results and commercial impact Oxford Instruments Analytical Oy (OIA) has developed GaAs detector technology in 3 successive ESA TRP and GSTP projects. In the first two projects the emphasis was on spectroscopy, but in the last one (2003 05) pixellated detectors enabling x-ray imaging were fabricated. The figure 3 shows an x-ray image obtained with a 150 µm thick GaAs detector which has been bump bonded to a Medipix-1 chip at VTT. Figure 2. A close up of the imaging matrix of 3 x 4 picels. In mammography applications, GaAs detectors will have small pixels, around 50 µm, and the active thickness (depleted layer of the pin diodes) of the detectors should be at least 100 µm thick, preferably 200 µm. GaAs detectors that have very small pixels, 25 µm x 50 µm were successfully produced in this project. Unfortunately, the thickness of the active layer only occasionally exceeded 100 µm due to variation in the quality of the GaAs wafers. As a consequence, a new project to improve the material quality has been started with the GaAs wafer provider. 65

Figure 3. X-ray image of a sardine, obtained with a GaAs detector flip-chip bonded to a Medipix-1 readout chip (http://medipix.web.cern.ch). The detector has 64x64=4096 pixels of the size of the detector, which is 170 µm x170 µm. A. Owens (ESA), H. Andersson (OIA), M. Campbell (CERN), D.Lumb (ESA), S. Nenonen (OIA), L. Tlustos (CERN), Proceedings of the SPIE, Volume 5501, pp. 241-248 (2004). Figure 4. Close-up of the GaAs detector processed at Modulight. The detectors are pin diodes and pixels were fabricated by etching mesa structure on the p-side. The pitch of the pixels is 25 µm x 50 µm. The diameter of the bump pads is 50 µm. The resistance between pixels was several Gohms. The size of the detector is 1 cm x1cm. After the material problem was solved, the development of the GaAs-based mammography detector system continued with the partners. Besides mammography applications, OIA has started to develop its GaAs detector technology for security and synchrotron applications in collaboration with foreign partners and end-users. The potential market for the planned security application can even exceed the mammography detector. VTT: results and commercial impact VTT is a leading supplier of fine-pitch bumping and flip-chip hybridization. VTT s task in this project was to develop a hybridization technique for detector assemblies consisting of GaAs pixel sensors and customized CMOS readout circuits. To realize the desired pixel size of 25 µm x 50 µm, a new 10-µm solder bumping process for 200-mm readout wafers was set up. Electroplated solder alloys were used to deposit the bumps. A new TiW/Cu field metal sputtering and a copper spray etching process were developed to enable repeatable bumping with high yield. At the same time processing tools were updated with improved sputtering and spray etching systems. The flip-chip bump bonding process was evaluated using pixellated GaAs sensor chips and bumped passive readout chips allowing characterization of individual pixels after hybridization. In addition to the eutectic SnPb solder alloy bumps, a low melting point (~ 90 C) SnPbBi alloy was also tested. A hybridization procedure maintaining the good electrical characteristics of the sensor and mechanical strength of the bonded assembly was found. An active Medipix1 readout wafer was bumped using eutectic SnPb bumps and chips were flip-chip bonded with matching GaAs sensor chips succesfully. Improving the quality of GaAs sensor chips was found to be the key issue in realizing bump bonded assemblies with high yield. 66

Figure 5. SEM image on an 10-µm bump array with nominal pixel size of 25 µm x 50 µm. Minimum pitch on the layout is 29 µm. The metallurgical structure (from substrate up) of bumps is sputtered TiW/Cu field metal stack, electroplated Ni, and electroplated eutectic SnPb solder alloy. A 10-µm solder bumping process and flip-chip bonding procedures compatible with GaAs pixel sensors were developed enabling VTT to make small-scale detector hybridization for the partners. Process improvements developed for the 10-µm process were introduced at VTT s multipurpose 30-µm bumping process to further improve the yield of the bumping process. Contact information Planmed Oy Pekka Strömmer, vice president Tel. +358 20 7795 525 pekka.strommer@planmed.com www.planmed.com Oxford Instruments Analytical Oy Seppo Nenonen, manager, space projects Tel. +358 9 32841 341 Seppo.Nenonen@oxinst.fi, www.oxinst.com VTT Jorma Salmi, senior research scientist Tel. +358 20 722 6639 Jorma.Salmi@vtt.fi www.vtt.fi 67

3.5 Finnish-Canadian cooperation projects Development and Demonstration of Remote Sensing Based Products For Surface Water Management (WRPD) The Consortium The Consortium consists of the Finnish Meteorological Institute (FMI), the Helsinki University of Technology (TKK), the Finnish Environment Institute (SYKE), and Affecto Genimap Oyj. Affecto Genimap is responsible for overall coordination of the project. The consortium members have the following responsibilities: Affecto Genimap Developing new channels and business models to deliver EO information from data providers to end users Creating data management methods from EOdata to end user products as well as combining it with European wide map content TKK Algorithm development for snow covered (SCA) area estimation using space-borne SAR Realization of an operative SCA processing chain incorporating Radarsat SCW data Development of SWE retrieval capabilities from spaceborne passive microwave sensors Development together with SYKE of a technique to produce water quality information from multi-temporal MODIS images. SYKE Algorithm development for estimating the snow covered (SCA) area from the optical EOdata Development of the operative system for EO-data acquiring, processing and delivery to the AffectoGenimap Adapting the operative system to process radar images Implementation of the Radarsat SCA algorithm provided by TKK to the operative system. Testing and implementation of the ice thickness and ice-break detection algorithms provided by Noetix Research to the operative system Implementation of the turbidity algorithm in the operative system Improvement of the quality of the end products. FMI Algorithm development for estimating the boreal forest leaf area index (LAI) from SAR data. Microwave radiometer snow product development in co-operation with TKK. Project duration: from 1 st October2004 to 31 st 2007 volume 534 k. Project volume: the volume of Finnish research part is 540k : 175 k for TKK, 257 k for SYKE and 108 k for FMI. Background AffectoGenimap AffectoGenimap, a mobile information systems provider in Finland, has developed the interfaces to deliver map data and aerial images from commercial databases to a broader user market. AffectoGenimap has recognized that EO-data will be a significant field for enlarging its product portfolio with on-line and mobile services. The goal is to combine snow, water quality/quantity, and ice information as the theme with other GIS information such as maps, road information etc. in order to visualize and locate the information more precisely. The other goal is to gain access to new markets and promote commercial exploitation and market development through corporate marketing initiatives FMI The WRDP project is continuing development of the LAI estimation algorithm based on SAR images. At the beginning of the project, the relationship between boreal forest LAI and VV/HH polarization backscattering ratio was demonstrated in small test areas. No experience of LAI retrieval using SAR data was available in Sub-Arctic areas. 68

TKK The WRDP project continues the development of satellite-based snow and water quality monitoring applications. With snow, the focus is on SAR-based SCA estimation using ERS, Envisat and Radarsat data and on estimation of SWE with SSM/I microwave radiometer data. The SCA method has been developed for the boreal forest region, and has now been altered for operational use. With water quality applications the focus is on using MODIS and MERIS data for the mapping of characteristics such as turbidity and the concentration of chlorophyll a. SYKE One of the main objects in the WRPD project is to create end-to-end capability from raw EO-data to value-added end product. This means that EOdata flows in near real time from the data provider through the SYKE operative system to Affecto- Genimap, where value-added end products are produced. The WRDP project also strongly promotes the research and development for the SYKE operative remote sensing applications. Research is conducted within the organisations and also with Finnish and Canadian partners. Project objectives 1. Develop advanced retrieval algorithms for each of the applications, 2. Develop an end-to-end capability to access earth observation (EO) products and deliver it to end-users. 3. Demonstrate the information integration and delivery to end-users. 4. Include the enhanced services in the European Space Agency (ESA) GMES (Global Monitoring for Environment and Security) Service Element (GSE), Polar View initiative. Activities AffectoGenimap The Genimap Platform was enhanced to 1) connect external data providers to deliver data from their internal servers to end users via Affecto- Genimap, 2) provide new interfaces to end users and applications, 3) manage EO-data. Mobile clients were developed to provide new user interfaces to use up-to-date EO-based information in wireless environment. Genimap Platform was enhanced with connector-technology to be ready for external data providers via standard interface. SYKE In the WRDP-project, SYKE has built an interface where interpreted EO-data are available in standardized format in near real time. Current applications include sea surface temperature (SST), surface algae intensity, turbidity, and optical snow covered area (SCA) well over 200 EO derived end products yearly. Besides operative tasks, research and development of the operative applications are conducted in the project. This includes algorithm development, validation, field surveys, and research aimed at accuracy improvements. The project also endorses co-operation with research partners, both national and international. At the project midpoint, national co-operation has resulted in the implementation of the TKK turbidity algorithm and Radarsat SCA process to the SYKE operative system. The test phase for the latter one starts in spring 2006. Research co-operation between SYKE and FMI includes transmissivity and leaf area index (LAI) comparisons aim at possible improvements in optical SCA estimation. SYKE has promoted co-operation with Canadian counterparts in 3 distinct areas: optical snow estimation, lake and river ice, and water quality. The SYKE SCAmod method was assessed across part of northern Manitoba, Canada. Canadian Noetix Research (CAN) demonstrates lake and river ice interpretation from radar images at two test sites in Finland. The demonstration will be presented to Finnish end users. Water quality co-operation seeks to demonstrate Finnish water quality algorithms in Lake Winnipeg, Canada. TKK An operative SCA processing chain incorporating Radarsat SCW data has been achieved. The 69

automatic SCA processing chain has been developed, validated and delivered to SYKE and will be implemented for operative demonstration use during spring 2006. There has further development of the SCA algorithm, and a wide range of analyses has been performed to validate the accuracy of the estimation method. To comply with SYKE s future EO products, a novel grid-based SCA estimation method has been developed. Development of SWE retrieval algorithms using passive microwave instruments has been on-going. The results have been demonstrated for northern Eurasia and Canada. An empirical method for producing turbidity maps of the Gulf of Finland using multitemporal MODIS data was developed and handed over to SYKE for implementation and validation. FMI The boreal forest LAI retrieval algorithm was further developed to be applicable to large areas. It was tested in 300 m and 1 km resolution in Central Finland and the overall agreement with corresponding LAI maps derived from a SPOT image was good. Development of an LAI estimation algorithm for Sub-Arctic areas was started. Research results The following includes the results achieved during the first year and half of the WRDP project (total 3 years). AffectoGenimap Enhancements in Genimap Platform in order to satisfy the needs for EO-data delivery New mobile/web/wms clients for EO-data delivery Develop managing & maintaining methods SYKE Creation of End-To-End capability from raw EO data provider via the SYKE operative system to end users Creation of new end product package including numerical data, thematic map and metada. New SCA product with improved resolution Implementation of the turbidity algorithm developed by TKK for the SYKE operative system Implementation of the Radarsat SCA estimation process developed by TKK for the SYKE operative system Conference papers to IGARSS05 and IGARSS06. Optical SCA demonstration with SYKE method in Manitoba area, Canada. Operative activities resulting in 235 end products during 2005 (partly supported by the project). TKK Development of an automatic processing chain for SCA estimation using Radarsat data Implementation of a grid-based SCA estimation method to insure compliance with future SYKE EO products Development of a turbidity algorithm for the Gulf of Finland Development and demonstration of an SWE mapping technique for boreal forest and sub-arctic zones. Conference papers to IGARSS05 and IGARSS06 Several articles in peer-reviewed journals. FMI Adaptation of the LAI estimation algorithm for large areas using SAR VV/HH images Conference paper to IGARSS05 Published paper in IEEE Transactions on Geoscience and Remote Sensing Summary of the LAI algorithm at the ESA ENVISAT AO project PI portal. Company results and commercial impact Providing up-to-date EO data via existing channel to new customers, AffectoGenimap believes that it can generate new business opportunities. Four potential end users were suggested to the project s steering group. This was done to give an accurate and realistic view of development from the end user s point of view. 70

Contact information AffectoGenimap Oyj, BI and GIS Solutions Esa Orava, team leader Tel. +358 40 591 2771 esa.orava@affectogenimap.fi www.affectogenimap.fi Finnish Environment Institute (SYKE), Geoinformatics and Land Use Division Saku Anttila, researcher Tel. +358 9 4030 0664 saku.anttila@ymparisto.fi www.ymparisto.fi Finnish Meteorological Institute (FMI), Earth Observation Terhikki Manninen, senior research scientist Tel. + 358 9 1929 4159 terhikki.manninen@fmi.fi www.fmi.fi Helsinki University of Technology (TKK), Laboratory of Space Technology Kari Luojus, research scientist Mobile +358 40 5058417 kari.luojus@tkk.fi P.O.Box 3000, FIN-02015 TKK, Finland www.tkk.fi of primary importance to the Canadian end users in the project. Arbonaut will calculate the forest biomass estimates. Arbonaut s focus is on quantitative forest analysis according to the priorities of the Finnish end users. Arbonaut will also develop software products based on the technology developed in the project. Metla is responsible for forest growth modelling. Growth modelling will be adapted to both Finnish and Canadian circumstances. Project team and management The project teams are being led by Dr. Udo Nielsen of Dendron, Dr. Tuomo Kauranne of Arbonaut and Prof. Tuula Nuutinen from Metla. The project managers in each organisation are Andy Welch, Jari Kinnunen and Reetta Lempinen, respectively. The user group includes representatives from national and state forest management authorities and several large forest companies. Project duration: from 15 th May 2005 to 31 st December 2007 Project volume: The volume of Finnish research part is 281 k Project description and background Northern Boreal Forest Information Products Based on Earth Observation Data The Consortium The structure, participants and responsibilities of the consortium are presented in figure 1. The consortium comprises the Finnish Forest Research Institute Metla, and the technology companies Dendron, based in Ottawa, and Arbonaut, based in Joensuu, Finland. Dendron is responsible for the collection and preprocessing of forest imagery. They will provide visual assessment and land use analysis, as well as all basic mapping products. Dendron s other main focus is ecological classification, which is The project aims at producing forest information service products for public and private customers, particularly targeting areas that are difficult to access by land and beyond the limit of commercial forestry. The information produced is based on SAR satellite images and mathematical models that are calibrated with data collected from densely observed areas. The products are envisaged to produce forest biomass estimates, basic mapping data, and an indicative ecological classification for forest stands. The international adoption of methods applied in MELA, though general as such, is hindered by the limited availability and/or compatibility of tree-level forest resource data and models outside Finland. On the whole, local stand-based approaches are still used; this means a lack of interoperability between the forest information sys- 71

CSA CLIENT CONTACTS TEKES Udo Nielsen Dentron PROJECT MANAGERS Tuomo Kauranne Arbonaut WP1 Image Analysis & Base Mapping TEAM LEADER Udo Nielsen, Dendron Partner(s): Rob Arnup WP2 Data Assimilation & Dissemination TEAM LEADER Tuomo Kauranne, Arbonaut WP3 LocalMELA TEAM LEADER Tuula Nuutinen,METLA Partner(s): University of Joensuu WP4 Commercialization TEAM LEADER T&D Finland UPM Finnish Forest and Park Service Metsämannut End Users Canada Ontario Ministry of Natural Resources Inventory Monitoring & Assessment Services Northern Boreal Initiative Figure 1. Participants and responsibilities of the consortium. tems of different companies and the providers of forest resource data: it has not been possible to incorporate information based on individual tree modelling in them. Moreover, the modification of the first software implementation of MELA simulation to local conditions was difficult and time-consuming, requiring lots of expertise because of the prevalent software technology. Project objectives WP2: To identify and test algorithms that are able to generalize the information obtained from densely observed areas guided by satellite observation into sparsely observed areas. If dense measurements are outdated, they will be brought up to date by growth modelling tested in WP3. WP 3: To modify Metla s extranet service for calculating forest growth estimates when assimilated boreal forest data and locally applicable models are available and stored using standard formats in the Laani database and model archive of the extranet service, respectively. Activities WP2: Densely observed calibration data have been collected and the first trial algorithms have been tried out on it. In the next stage, this process will be extended to the analysis of different forms of satellite data, regarding the appropriateness for spatial generalization. WP3: Metla s extranet service is based on three components: 1. an application interface tailored for the specific client access in this project 2. a specialised algorithm for the calculation of model-based characteristics such as forest growth. The algorithm calculates the results at calculation unit level (either sample plots or stands) based on tree-level models. When individual tree measurements are not available, alternative calculation chains, based on synthetic tree distributions, will be applied 3. a model archive where locally applicable models are stored for two sites (one in Finland and one in Canada). Expected results WP2: Effective spatial and temporal generalization algorithms have been identified, tried and verified. WP3: a. An extranet service for calculating forest growth estimates based on assimilated boreal 72

forest data and growth models defined by local experts and stored in the model archive in standard interface format. b. A new growth model interface standard that makes it possible to extend the service for different conditions with the help of local experts Company results and commercial impact To be decided in WP4 (during 2007). The project is still underway. Contact information Finnish Forest Research Institute, Joensuu Research Unit Tuula Nuutinen, professor Tel. +358 10 2111 Tuula.Nuutinen@metla.fi http://www.metla.fi/metinfo/mela/index.htm Oy Arbonaut Ltd Tuomo Kauranne, president Tel. +358 13 259 1911 tuomo.kauranne@arbonaut.com www.arbonaut.com Applications and software for SAR interferometry and coherent target monitoring (AppliSARIN) The Consortium The APPLISARIN team is made up of organisations from both Canada and Finland. The team consists of Vexcel Canada Inc.; the Canadian Forest Service, National Resources Canada; the Finnish Geodetic Institute; the Helsinki University of Technology; and Soil and Water Oy. Each team member brings a particular expertise to the consortium. Vexcel develops and markets commercial Earth Observation software products and provides services to users around the world. The Canadian Forest Service and the Helsinki University of Technology are at the forefront of research for forestry applications. The Finnish Geodetic Institute has extensive experience in the field of surface deformation and Soil and Water Oy are EO-data and service providers. There are end-users for the APPLISARIN project in each of the three areas of interest. For surface deformation, the end-users are Posiva Oy; Helsinki City, Geotechnical division; Turku City, Real Estate Department; and, the Vexcel EOSserv Department. The Canadian Forest Service is an end-user in forestry and will provide contact with users in provincial governments and private industries. Kemira GrowHow Ltd will provide precision agriculture services using aerial images and is interested in the ability of radar satellites to provide this information. Project duration: from 1 st October 2004 to 31 st December 2006. Project volume: The volume of Finnish research part is 455 k : 300 k for FGI and 155 k for TKK Background With traditional differential SAR interferometry (DINSAR) using two to four SAR images, height accuracies of a few meters and sub-cm surface deformation measurement accuracy are feasible under favourable conditions (Bürgmann et al. 2000). With interferogram stacking techniques, even more accurate results can be obtained to increase fringe clarity and to decrease errors due to atmospheric delay (Sandwell et al. 1998). Temporal and geometric decorrelation and atmospheric inhomogeneities often corrupt the interferogram and affect the accuracy of the results. To overcome atmospheric inhomogeneities and temporal and geometrical decorrelation, a procedure has been developed for identifying and exploiting stable natural reflectors. This procedure, often referred to as permanent scatterers (PS) or coherent target monitoring (CTM), uses a long temporal series of interferometric SAR images (Ferretti et al. 2000, 2001). These techniques have been used for land subsidence and known seismic fault monitoring in urban areas (Ferretti et al. 2001, 2000; Colesanti et al. 2002), and ground deformation measurement in rocky/mountainous areas (Dehls et al. 2002. Presently, there is a need to develop practical methods related to DINSAR for surface deformations. Forest stem volume is one the key characteristics in forest inventories. Satellite sensors have proved useful for mapping forest stem volume for 73

large areas, but their accuracy for small areas, such as for individual forest stands (size 1 3 ha), has been inadequate. The performance of the ERS-1/2 Tandem mission has been relatively good, almost the same as that obtained with optical satellite imagery (e.g. Hyyppä et al. 2000). Temporal decorrelation is the major cause of inaccuracy with biomass estimation using INSAR. Single-pass satellite interferometry or repeatpass interferometry with small temporal delay may lead to improved results. The ENVISAT and ERS-2 satellites have different centre frequencies and are currently in orbits that have them spaced ½ hour apart with a baseline of up to 2,000 meters. This small temporal delay between the satellites is expected to reduce the decorrelation and thus improve the biomass estimation. In agricultural monitoring, satellite images are needed regularly and the time window for the best acquisition date is narrow. Agriculture has been identified as one of the most important potential business areas for SAR images in the future (ERSIS report, 2000). The estimation of the vegetation biomass from SAR images is difficult and one of the best methods has been the use of ERS-1/2 1-day tandem coherence information. There have been promising results from correlating interferometric coherence with crop biomass (Blaes et.al., 2003). It is expected that the ERS- ENVISAT satellite combinations will provide similar results to those obtained with the ERS-1/2 tandem data. Project objectives The objective of the APPLISARIN project is to create proven monitoring tools for surface deformation, forest biomass estimation, and crop monitoring using new methods and software for SAR interferometry enabling ENVISAT/ERS-2 interferometry, and enabling alternating polarisation of ENVISAT INSAR and Coherent Target Monitoring. Research results Some examples of the project research are shown below. The techniques based on stable targets are needed to measure deformation in areas where traditional DINSAR measurements are not feasible. The plentiful vegetation in southern Finland decreases coherence and makes interferograms very noisy. According to our studies, however, deformation can be extracted with long time series of data and advanced processing algorithms. Spatially extensive data regarding the subsidence of buildings in city areas was obtained and even millimeter scale movement can be detected. The rates of subsidence agreed well with the reference data, but since the accuracy of reference measurements was unknown, but believed to be roughly one millimeter, this can be considered a very good result. However, there are some limitations to the application of the advanced DINSAR techniques. The size of the area of interest must be small enough. The density of stable targets is also an important parameter. The coherent target approach does not work for a sparse network of persistent scatterers. Another algorithm, for example the permanent scatterers algorithm, which exploits different criteria for persistent scatterer selection, might be more appropriate for areas having a sparse network of PSs. A high coherence non-subsiding area is needed for atmospheric phase determination. The selection of atmospheric correction reference area is of high importance and affects the results dramatically. The comparison of the deformation rates from the two data sources is not straightforward. In our case, the levelling points and the INSAR coherent targets were situated in different parts of the building. The levelling benchmarks were in the stone foundations of the buildings. On the other hand, the INSAR coherent targets were more likely in the roof structures and often only on one side of the building. Determining the corresponding measurement points had to be done manually. In addition, the different coordinate systems of the data sets made the overlay process difficult and prone to errors. Nevertheless, the subsidence rates from the two different sources were consistent. The first spatially extensive measurements from Turku and Helsinki were obtained with advanced INSAR technique. Our results demonstrated that the INSAR measurements can provide a good overall picture of subsidence in these cities. Anyhow, to obtain more detailed deformation about 74

Figure 1. An INSAR deformation map of Helsinki, in which two subsidence areas are highlighted: Rautatientori and Jätkäsaari. building subsidence, levelling measurements are still needed, since the coherent targets are not usually evenly distributed in the buildings. Research concerning forest biomass estimation with a literature survey, which is appended to this document. Suitable models for forest biomass estimation were selected according to this survey for testing. These models are the TKK empirical coherence forest model and the Interferometric Water Cloud Model. The performance of models was tested with old ERS-1/2 Tandem data and ERS/Envisat coherence pairs if available. Also, a simple exponential model was tested. The good property of this simple model is that it does not require detailed ground truth information; some simple threshold values for coherence and maximum biomass are sufficient. At this moment, the implementation of models using Matlab is still under way. The purpose of Vexcel is to implement the chosen estimation algorithm in their SAR-processing software. In agricultural remote sensing, applied images should be acquired frequently enough to monitor important crop growth stages. Thanks to the cloud penetrating and flexible swath-positioning capabilities of space-borne SAR, at present images can be acquired even at the interval of few days during a growing season. In this study, dual-polarization (VV/VH) Envisat SAR images were used in association with practical ancillary data to monitor crop growth and to classify crop species. It was noticed that the high temporal resolution enabled nearly continuous monitoring, but it also caused problems because of the varying incidence angles. Moreover, the conduction of field surveys rapidly enough for research purposes was observed to be a problem. R 2 of 0.55 was obtained for estimating the crop growth, when average crop height in parcels was used to describe the amount of biomass. Overall accuracy of 74.7% was achieved for crop species classification. VH polarization appeared to be useful in the estimation, even though the noise equivalent ó 0 was too high to detect early crop growth. Field based averaging was required; hence for precision farming purposes a better spatial resolution would be needed to detect biomass variations within parcels. 75

Contact information FGI Juha Hyyppä, professor Tel. +358 9 295 550 juha.hyyppa@fgi.fi www.fgi.fi TKK Henrik Haggrén, professor Tel. +358 9 4511 hannu.hyyppa@hut.fi www.foto.hut.fi Soil and Water Oy Miranda Saarentaus, director Tel. +358 9 682 6531 Miranda.Saarentaus@poyry.fi www.poyry.fi Multi-Polarisation SAR for Operational Sea Ice Monitoring (POL-ICE) The consortium The project is structured into two parallel strands: one in Canada tailored towards the requirements of the Canadian end-user, the Canadian Ice Service, and one in Finland tailored towards the requirements of the Finnish end-user, Finstaship, which operate icebreakers for the Finnish government. Each strand shares the same overall objectives and follows the same main project stages. Regular interaction is planned to ensure technology transfer through exchange of data, results, and technical workshops. In Finland, the project leader is the Laboratory of Space Technology / Helsinki University of Technology (TKK). The Partners are the Finnish Institute of Marine Research (FIMR), VTT Information Technology (VTT) and Finstaship, which is also end-user. In Canada the project leader is MacDonald Dettwiler and the partners are Noetix Research Inc, RADARSAT International (RSI), University of Calgary (UC) / Department of Geography. The end-user is the Canadian Ice Service. POL-ICE comprises seven main tasks, some of which are divided into several sub-tasks. Throughout the project, the partners work very closely with each other; none of the main tasks will be conducted by only one project partner. Hence, all the project progress reports and most of reports detailing project results are being written together. Responsibilities of the Finnish partners HUT is the project leader and maintains an extensive web-site on the project: http://www.space.tkk.fi/research/projects/ pol-ice/index.html HUT participates in the development classification algorithms for the SAR images of sea ice and conducts basic sea ice research supporting this task. HUT and FIMR will conduct radar and ground truth measurements during the POL-ICE field campaign in 2007. The field campaign data will support and validate development of the SAR classification algorithms. All the Finnish partners together will review the end-users requirements in sea ice monitoring and their opinions of current sea ice products several times during the project. This work will guide development of the sea ice products and end-user software for visualization and analysis. FIMR s main task is to develop SAR classification algorithms. FIMR and VTT together will study end-user data delivery issues. VTT s main task is to develop a pilot software demonstrator. The current end-user software will be enhanced to provide the user with tools that help to interpret the current and new sea ice products. VTT together with FIMR and Finstaship are responsible for end-user (Finstaship ice breakers) training and feedback related to the pilot demonstrator. Finally, FIMR and VTT will study commercialization aspects of the sea ice products. Project duration: from 1 st January 2005 to 31 st December 2007 Project volume: The volume of Finnish research part is 618 k ; 158 k to TKK, 275 k to FIMR and 185 k to VTT 76

Project description Timely and accurate information on sea ice is of critical importance to the economies of Canada and Finland. Both countries maintain national ice services with responsibility for generating daily ice charts and forecasts and disseminating these to coastguard, icebreakers and other maritime operators. Space-borne synthetic aperture radar (SAR) is the primary data source for operational sea ice monitoring. Until recently, only singlepolarised SAR data (ERS-1/2, RADARSAT-1) was regularly available. Research has shown that considerably more information on sea ice characteristics (e.g., ice type classification) can be extracted from SAR when multiple polarisations are available. The increased information content also improves the potential for automated analysis. Such data is becoming routinely available with the alternating polarisation mode of ENVISAT ASAR launched in 2002. RADARSAT-2, which is to be launched in late 2006, will provide dual-polarisation with all RADARSAT-1 heritage modes and fully polarimetric data over narrow swaths. With this wider availability of multi-polarisation SAR data, there is now a need to transfer research on sea ice monitoring applications to the operational domain. To address this need this proposal brings together Canadian and Finnish experts in sea ice monitoring and multi-polarisation SAR with the following objectives: To develop demonstrations of the use of dual-polarisation SAR data from ENVISAT ASAR and RADARSAT-2 and other ancillary data sources to improve operational sea ice monitoring in both Canada and Finland. To demonstrate the use of fully polarimetric SAR data from RADARSAT-2 for enhanced tactical sea ice classification in areas of critical interest. To facilitate transfer of new sea ice monitoring algorithms and image analysis methods between Canada and Finland. To identify how the new algorithms and image analysis methods can be best integrated into existing sea ice monitoring operations. To expand the market for ENVISAT and RADARSAT-2 data by operationally demonstrating new SAR derived sea ice monitoring products and services and promoting these to potential end-users. To provide more efficient and safer navigation in sea ice affected areas. The main project tasks in Finland are 1) a review of sea ice monitoring requirements, 2) algorithm development, 3) data delivery issues, 4) pilot demonstrator development, 5) fieldwork validation, 6) end user training and feedback, and 7) commercialization. Research results The three-year POL-ICE project has been running for only one year. The project work during the first year has mainly concentrated on the task Review of sea ice monitoring requirements. Additionally, some work has been conducted in algorithm development and pilot demonstrator development. The algorithm development work has been focused to add information about the degree of ice deformation to the operative sea ice thickness chart product. In 2005, two end-user seminars were held and a feedback form was sent to Finnish ice breakers. The seminar topics included formats of the sea ice products, new features of end user-software and analysis of the feedback forms. The results of the seminars will guide project development work. New seminars and end-user inquiries will be held in 2006 and 2007. The first Canada-Finland POL-ICE workshop was held on 1 November in Ottawa, Canada. In the workshops participants presented their previous and current sea ice research and discussed the results of end-used requirement reviews and synergies between Canadian and Finnish work. Company results and commercial impact Wintertime ship traffic in the Baltic Sea is significant and important for many European countries, e.g. 40% of the annual 90 million tons of cargo transported by sea to Finland takes place during the ice season. It corresponds to tens of thousands of port-calls. During the last ten years, ship traffic to Finland has increased by 30%, and this trend is 77

Figure 1. Visualisation of co- and cross-polarized ASAR images using colours. The HH- and HV polarised images to the left are combined into a false coloured image to the right, thus enabling the user (icebreaker) to make a better interpretation of the image and to find the easiest way through the ice. Different visualisation techniques are part of the functionalities that will be tried out in the POL-ICE project. expected to continue. However, after Russia completes its large oil terminals near Vyborg, marine traffic growth is expected to be much faster in the Gulf of Finland. In 2001, there were about 38,000 port calls in the area; this is expected to increase to 52,000 in 2015. In 2002 some 76 million tons of oil was transported; by the end of 2015, oil transportation is expected to be more than 200 million tons. Consequently, there is increasing demand in the Baltic Sea for high-resolution sea ice data and information products with good spatial and temporal coverage. Sea ice thickness is an important part of this information. Considerable economic savings can be achieved by optimising ice navigation and reducing sailing times of ships. The planned new products will be a part of the Finnish Ice Service s strategy to develop information products that correspond to user requirements for NRT products at the user s scale and that are easy to interpret. Products developed in the project will be delivered to Finstaship icebreakers for end user validation. The Ice Service plans to launch the products on the shipping market sector when they are mature. As it is a leading ice service in the Baltic Sea area, the Finnish Ice Service has a strong position in the region and very good contacts to users. FIMR also chairs the Baltic Sea Ice Meeting, the Baltic Sea ice service and icebreaking co-operation working group. Thus the Finnish Ice Service is ideally placed for distributing information gathered during the project and distributing it to existing and new public and private customers. Contact information Helsinki University of Technology, Laboratory of Space Technology, Martti Hallikainen, professor Tel. +358 9 451 2371 martti.hallikanen@hut.fi Project www-site: http://www.space.tkk.fi/ research/projects/pol-ice/index.html 78

Attachments Companies 3D Systems Oy Aboa Space Research Oy Affecto Genimap Oy Arbonaut Oy Centroid Oy Componeering Inc. Elektrobit Microwave Oy Elisa Oyj European Communications Engineering Oy Fastrax Oy Fifth Element Oy FM-Kartta Oy INPHO Technology Oy Insinööritoimisto Toikka Oy Keyfor Oy/Keypro Oy Modulight Oy Nemo Technologies Oy Nokia Oyj Patria Systems Oy PIEneering Oy Pivotal Consulting Oy Planmed Oy Savcor Indufor Oy Shipping Enterprise - Fintaship Sito Oy Soil and Water Oy, Jaakko Pöyry Infra Space Systems Finland Oy SPECIM, Spectral Imaging Oy Stora Enso Oyj Terrasolid Oy Toiminimi Timo Pekkonen Turku Science Park Oy u-nav Microelectronics Finland Oy Vaisala Oyj WM-Data Oy Research and other organisations City of Helsinki Finnish Athletics Organization (SUL) Finnish Environment Institute (SYKE) Finnish Forest Research Institute (Metla) Finnish Geodetic Institute (FGI) Finnish Institute of Marine Research (FIMR) Finnish Maritime Administration (FMA) Finnish Meteorological Institute (FMI) Finnish Road Administration Finnish Road Enterprise Helsinki Metropolitan Area Council (YTV) Helsinki University of Technology (TKK) Institut Cartogràfic de Catalunya (ICC) Millimetre Wave Laboratory of Finland (MilliLab) National Land Survey of Finland National Research Institute for Earth Science and Disaster Prevention (NIED, Japan), Radiation and Nuclear Safety Authority of Finland (STUK) Shipping Enterprise - Finstaship Stok - Development Centre Tampere University of Technology (TUT) Technical Research Centre of Finland (VTT) University of Helsinki (UH) 79

Tekes Technology Programme Reports in English 8/2006 AVALI Business Opportunities from Space Technology 2002 2005. Final Report. 79 p. 6/2006 New Knowledge and Competence for Technology and Innovation Policies ProACT Research Programme 2001 2005. Final Report. Edited by Pekka Pesonen. 137 p. 3/2006 ELMO Miniaturising Electronics 2002 2005. Final Report. 238 p. 12/2005 NETS Networks of the Future 2001 200. Evaluation Report, Executive Summary. 19 p. Mervi Rajahonka and Mikko Valtakari. 1/2005 NETS Networks of the Future 2001 2005. Final Report. 213 p. 10/2004 Competitiveness through internationalisation Evaluation of the means and mechanisms for promoting internationalisation in technology programmes. Evaluation Report. 89 p. Kimmo Halme, Sami Kanninen, Tarmo Lemola, Erkko Autio, Erik Arnold, Jesper Deuten. 6/2004 Developing Technology for Large-Scale Production of Forest Chips Wood Energy Technology Programme 1999 2003. Final Report. 98 p. Pentti Hakkila. 22/2003 Presto future products. Added Value with Micro and Precision Technology 1999 2002. Final Report. 110 p. 21/2003 Evaluation of the Finnish-Swedish R&D programme EXSITE, 2001 2003, Evaluation Report. 73 p. Risto Louhenperä, Olle Nilsson. 13/2003 Targeted Technology Programmes: A Conceptual Evaluation Evaluation of Kenno, Plastic processing and Pigments technology programmes. Evaluation Report. 104 p. Erkko Autio, Sami Kanninen, Bill Wicksteed. 10/2003 VÄRE Control of Vibration and Sound Technology Programme 1999 2002. Final Report. 90 p. 6/2003 Towards a competitive cluter An evaluation real estate and construction technology programmes. Evaluation Report. 89 p. Petri Uusikylä, Ville Valovirta, Risto Karinen, Enno Abel and Thomas Froese 5/2003 Developing technology for large-scale production of forest chips. Wood Energy Technology Programme 1999 2003. Interim Report. 53 p. 4/2003 Code Technology Programme 1999 2002. Final Report. 1/2003 FFusion 2 Technology Programme 1999 2002. Final Report. 151 p. Seppo Karttunen, Karin Rantamäki (eds.) Subscriptions: www.tekes.fi/english/publications

AVALI Business Opportunities from Space Technology Final Report The Finnish Funding Agency for Technology and Innovation Kyllikinportti 2, P.O. Box 69, FIN-00101 Helsinki, Finland Tel. +358 1060 55000, Fax +358 9 694 9196, E-mail: tekes@tekes.fi www.tekes.fi March 2006 ISSN 1239-1336 ISBN 952-457-230-3